Support for lexicography


I predict that when collaboration tools drop in Anytype there will be a storm of confusion as people who have each built slightly (or greatly) different ideas of the same things discover that knitting them all together is a major undertaking.

By planning to support lexicography, the practice of compiling dictionaries or lexicons, Anytype can provide effective tools to help users negotiate agreements across multiple spaces, especially when changes to a space based on another space need careful planning.


Fortunately, Wikidata has already built an excellent foundation in order to support Wiktionary. Its types and relations could be reimplemented in Anytype, with additional relations and UI to support easy creation and management of lexemes and lexicons.

The key is to ensure that the UI for lexicographical types and relations is geared for maximum convenience. This is because users will almost always create dictionaries to support each other’s mutual understanding, not because the dictionaries are inherently valuable. When users go to create, read, edit, or delete a lexeme, they will be doing so as part of another, more interesting task.

From this point, I will assume that we implement the types Lexicon, Lexeme, Form, and Sense, together with appropriate relations. In particular, lexemes will belong to at least one lexicon, but forms and senses will have independent existence, and any number of lexemes in number of lexicons may link to them, including across space boundaries. In this way, forms and senses will be the glue users use to bridge (or separate) their lexemes and lexicons

The most important component of this would be what I’m calling supertext. Supertext is hypertext, but it includes a separate “layer” of lexical hyperlinks, or lexlinks. A lexlink links a word, phrase, or symbol to its sense in a lexicon. Lexlinks enable authors to explicitly specify the senses of their written words and symbols, and they do not interfere with normal hyperlinks, which take precedence except when users specifically indicate that they want to follow a lexink (e.g. by holding down a modifier key or longpressing on a touch display). Authors can add lexlinks however they please, such as to entire objects or blocks, to spans of text, to individual words and symbols or characters, or not at all.

Another component would be added procedures to ease dictionary management, such as commands to copy or move lexemes between different lexicons and to authors to “stack”, so that it is easy to first choose senses from a specific lexicon (e.g. The Jargon File) and then from a general lexicon (e.g. The Oxford English Dictionary).


Supertext would make it possible for text in blocks in objects from different people and spaces to explicitly address exactly the definition intended, thus removing the need for users to manually search and guess the correct sense in each other’s personal dictionaries. This could be of great help to lawyers, software engineers, and writers of all kinds who need to take the greatest care that their words say what they mean while working together.

Lexicons could created for different purposes. One lexicon might capture a specific person’s personal language, while another might capture the language of a specific document, and another still might capture the language of a small community, like a group of friends or a team in a larger organization.

Given enough time and data, large corpuses of supertext and the lexicons backing them could be of great value to research, ranging from history to artificial intelligence.


It’s not necessarily the case that a complete implementation in Anytype would be the best route. Actual professional lexicographers have long embraced electronic processes. It may be that that Anytype merely needs to support linking to existing software or file formats.