Hypothes.is as a Digital Zettelkasten for Neologism and Word Collection for Wordnik
In the book The Professor and the Madman: A Tale of Murder, Insanity, and the Making of the Oxford English Dictionary, Simon Winchester describes the pigeonhole and slip system that professor James Murray used to create the Oxford English Dictionary. The editors essentially put out a call to readers to note down interesting every day words they found in their reading along with examples sentences and references. They then collected these words alphabetically into pigeonholes and from here were able to collectively compile their magisterial dictionary. Those who are fans of the various methods of knowledge collection and management represented by the index card-based commonplace book or the zettelkasten, will appreciate this scheme as a method of collectively finding and collating knowledge. It’s akin to Paul Otlet and Henri La Fontaine’s work on creating the Mundaneum, but focused on the niche area of lexicography and historical linguistics.
The Professor and the Madman is broadly the fascinating story of Dr. W. C. Minor, an insane asylum patient, who saw the call to collect words and sentences began a written correspondence with James Murray by sending in over ten thousand slips with words from his personal reading.
Wordnik and Hypothes.is
A similar word collecting scheme is currently happening on the internet now, though perhaps with a bit more focus on interesting neologisms (and hopefully without me being cast as an insane asylum patient.) The lovely folks at the online dictionary Wordnik have been using the digital annotation tool Hypothes.is to collect examples of words as they happen in the wild. One can create a free account on the Hypothes.is service and quickly and easily begin collecting words for the effort by highlighting example sentences and tagging with “wordnik” and “hw-[InsertFoundWordHere]”.
So for example, this morning I was reading about the clever new animations in the language app Duolingo and came across a curious new word (at least to me): viseme.
To create accurate animations, we generate the speech, run it through our in-house speech recognition and pronunciation models, and get the timing for each word and phoneme (speech sound). Each sound is mapped onto a visual representation, or viseme, in a set we designed based on linguistic features.
So I clicked on my handy browser extension for Hypothes.is, highlighted the sentence with a bit of context, and tagged it with “wordnik” and “hw-viseme”. The “hw-” prefix ostensibly means “head word” which is how lexicographers refer to the words you see defined in dictionaries.
Then the fine folks at Wordnik are able to access the public annotations matching the tag Wordnik, and use Hypothes.is’ API to pull in the collections of new words for inclusion into their ever-growing corpus.
Since I’ve collected interesting new words and neologisms for ages anyway, this has been a quick and easy method of helping out other like minded word collectors along the way. In addition to the ability to help out others, a side benefit of the process is that the collected words are all publicly available for reading and using in daily life! You can not only find the public page for Wordnik words on Hypothes.is, but you can subscribe to it via RSS to see all the clever and interesting neologisms appearing in the English language as collected in real time! So if you’re the sort who enjoys touting new words at cocktail parties, a rabid cruciverbalist who refuses to be stumped by this week’s puzzle, or a budding lexicographer yourself, you’ve now got a fantastic new resource! I’ve found it to be far more entertaining and intriguing than any ten other word-of-the-day efforts I’ve seen in published or internet form.
If you like, there’s also a special Hypothes.is group you can apply to join to more easily aid in the effort. Want to know more about Wordnik and their mission, check out their informative Kickstarter page.
A Small 10,000+ Annotations Party
If I’m honest, I get so much value and joy out of annotating online, I should have been the one to send them the care package.
Thanks Hypothes.is! I’ll see you in the margins.
Hypothes.is has methods for establishing document equivalency which archive.org apparently conforms. I did an academic experiment a few years back with an NYT article about books where you’ll see equivalent annotations on the original, the archived version, and a copy on my own site that has a
rel="canonical" link back to the original as well:
I don’t recommend doing the rel-canonical trick on your own site frequently as I have noticed a bug, which I don’t think has been fixed.
The careful technologist with one tool or another, will see that I and a couple others have been occasionally delving into the archive and annotating Manfred Kuehn’s work. (I see at least one annotation from 2016, which was probably native on his original site before it was shut down in 2018.) I’ve found some great gems and leads into some historical work from his old site. In particular, he’s got some translations from German texts that I haven’t seen in other places.
How to Make Notes and Write, a handbook by Dan Allosso and S.F. Allosso
A new handbook on note making and writing
I wasn’t expecting it until next week or shortly thereafter, but just in time for the new academic year, Dan Allosso has finished a major rewrite on his and S.F. Allosso’s earlier edition of A Short Handbook for writing essays in the Humanities and Social Sciences. This expanded edition has several new chapters on note making (notice that this is dramatically different than note taking) using a zettelkasten-based (or card index or fichier boîte if you prefer) approach similar to that practiced by Beatrice Webb, Marcel Mauss, Claude Lévi-Strauss, Roland Barthes, Hans Blumenberg, Mortimer J. Adler, and Walter Benjamin among many others.
The focus of the book is on note making for actively producing tangible outputs (essays, papers, theses, monographs, books, etc.), something on which a few recent texts in a the related productivity space haven’t delivered. While ostensibly focused on the humanities and social sciences in terms of examples, the methods broadly apply to all fields. In fact, some of the methods draw historically on some of the practices fruitfully used by Bacon, Newton, Leibnitz, Linnaeus, and many others in the sciences since.
This isn’t your father’s note making system…†
While many students (especially undergraduates and graduate students) may eschew this sort of handbook as something they think they “already know”, I can assure you that they do not and will benefit from the advice contained therein, particularly the first half. I’ve often heartily recommended Sönke Ahrens’ book How to Take Smart Notes: One Simple Technique to Boost Writing, Learning and Thinking to many in the past, but I think Allosso’s version, while similar in many respects, is clearer, shorter, and likely more easily realized by new practitioners.
There’s more detail in Dr. Allosso’s announcement video:
How to Make Notes and Write is available at Minnesota State’s Pressbooks site for reading online, or download as a .pdf or .epub. If you’d like a physical copy, they’re also available for purchase on Amazon.
For those in the educational spaces, Dr. Allosso has given the book a Creative Commons license (CC BY-NC-SA 4.0), so that people can use it as an Open Educational Resource (OER) in their classes and work.
For teachers who are using social annotation with tools like Hypothes.is in their classrooms, Allosso’s book is an excellent resource for what students can actively do with all those annotations once they’ve made them. (Here’s a link to my annotated copy of a recent working draft if you care to “play along”.)
† Unless of course your father happens to be Salvatore Allosso, but even then…
Hypothes.is, a web annotation tool, as an off-label zettelkasten?!
The platform has most of the basic functionalities one would want for a digital ZK:
- Simple note taking,
- Notes are editable and update-able,
- There’s a tag functionality for indexing notes,
- One can add links to other notes to cross link them if they liked,
- There’s a reasonably good search functionality,
- Data export is available if you want to move
The interesting piece is that if many are doing this in public, then folks can reply to others’ notes and even cross-link their public notes. (They’ve got the ability to have public and private notes, as well as groups for collaboration, which are functionalities most ZK don’t have as well.)
I was reminded of this potential off-label use case when someone replied to an older note with a quote/comment and it nudged me to add a note to link the two together. Who’s up for a public, social zettelkasten practice?
For those who need an example to look at, try my “digital ZK”: https://hypothes.is/users/chrisaldrich or their public timeline: https://hypothes.is/search.
I know some have mentioned Hypothes.is as an annotation tool for their note taking before. There are a few who use it as an online platform for notes and then they leverage the platform’s API or feeds to export their data to their tool of choice. (I’ve used Hypothes.idian for Obsidian to do just this.)
I get @TomCritchlow’s sentiment, but… the extra “work” it currently entails for the social part dramatically ups the signal to noise ratio for me compared to Twitter.
You’d definitely want the ability to filter by your social circle, especially on popular sites. In fact this sort of discovery mechanism would be cool if it could be more broadly built into either the web or perhaps into IndieWeb social readers which would know your social graph and could surface related details.
Perhaps expanding a browser extension like Crowdwise to include Twitter support might be a potential solution? I would worry that portions wouldn’t add much other than a lot of likes and bookmark-like data. While some,  might consider Twitter as an annotation layer (not always directly linked) on the web, the overall quality isn’t necessarily going to be built in there.
It would be cool if Crowdwise also added Hypothes.is’ API to their list of sources.
I’m also reminded of Peter Hagen’s experiments with Hypothes.is seem very similar but with a different UI. His version flips the discovery question on its head.
This is very similar to the sort of sensemaking and interlinking of information that Sönke Ahrens outlines in his book How to Take Smart Notes though his broader note taking thesis goes a few additional steps for more broadly synthesizing ideas into longer papers, articles, theses, and books.
Dr. Moskell also outlined a similar tactic at the Hypothesis Social Learning Summit: Spotlight on Social Reading & Social Annotation earlier today, though that video may not be accessible for a bit.
How can we better center and model these educational practices in our pedagogies?
Creating a commonplace book or zettelkasten index from Hypothes.is tags
I find that indexed subject headings can be useful for creating links between my wiki-like pages as well as links between atomic ideas in my digital zettelkasten. Gradually as one’s zettelkasten becomes larger and one works with it more, it becomes easier to recall individual ideas and cross link them. Until this happens or for smaller zettelkasten it can be useful to cross reference subject headings from one zettel to see what those link to and use those as a way to potential create links to other zettels. This method can also be used as a search/discovery aide for connecting edge ideas in new areas to pre-existing portions of one’s zettelkasten as well. Of course at massive scale with decades of work, I suspect this index will have increased value as well.
I don’t hear people talking about these types of indices for their zettelkasten in any of the influencer spaces or on social media. Are people simply skipping this valuable tool? For those enamored of Niklas Luhmann, we should mention that having and maintaining a subject index was a powerful portion of his system, even if the digitized version of his zettelkasten hasn’t yet been fully digitized. I haven’t seen the whole collection myself, but based on the condition of some of the cards in his index, Luhmann heavily used his subject index. (Note to self: I wonder what his whole system would look like in Obsidian?) Having a general key word/subject heading/topic heading index of all the material in one’s system can be very useful for general search and discovery as well. This is one of the reasons that John Locke wrote about a system for indexing one’s commonplace book in 1685. His work here is likely the distal reason Luhmann had one in his system.
Systems that have graphical knowledge graphs may make this process easier as one can look from one zettel out one or two levels to see where those link to.
Since such a large swath of my note taking practice starts by using Hypothes.is as my tool of choice, I’m able to leverage several years of using it to my benefit. Within it I’ve got 9,314 annotations, highlights, and bookmarks tagged with over 3,326 subject headings as of this writing.
To get all my subject heading tags, I used Jon Udell’s excellent facet tool to go to the tag editing interface. There I entered a “max” number larger than my total number of annotations and left the “tag” field empty to have it return the entire list of my tags. I was then able to edit a few of them to concatenate duplicates, fix misspellings, and remove some spurious tags.
An alternate way of doing this is to use a method described in this GitHub issue which shows how to get the tags out of local storage in your web browser. Your mileage may vary though if you use Hypothes.is in multiple browsers, which I do.
I moved this list from the tag editor into a spreadsheet software to massage the list a bit, clean up any character encodings, and then spit out a list of [[wikilinked]] index keywords. I then cut and pasted it into my notebook and threw in some alphabetical headings so that I could more easily jump around the list.
Now I’ve got an excellent tool and interface for more easily searching and browsing the various areas of my multi-purpose digital notebook.
I’m sure there are other methods within various tools of doing this, including searching all files and cutting and pasting those into a page, though in my case this doesn’t capture non-existing files. One might also try a search for a regex phrase like:
/(?:(?:(?:<([^ ]+)(?:.*)>)\[\[(?:<\/\1>))|(?:\[\[))(?:(?:(?:<([^ ]+)(?:.*)>)(.+?)(?:<\/\2>))|(.+?))(?:(?:(?:<([^ ]+)(?:.*)>)\]\](?:<\/\5>))|(?:\]\]))/ (found here) or something as simple as
/\[\[.*\]\]/ though in my case they don’t quite return what I really want or need.
I’ll likely keep using more local search and discovery, but perhaps having a centralized store of subject headings will offer some more interesting affordances for search and browsing?
Have you created an index for your system? How did you do it?
Unable to search or find public replies to annotations in public stream
Replies (with or without tags) to primary/original annotations are unable to subsequently be found in the main public stream or via search at https://hypothes.is/search.
Steps to reproduce
- Make a reply to any public annotation (with or without tags)
- Use https://hypothes.is/search to search the username of the reply or one of the original tags
- The reply can’t be found
The original (more complicated) example that uncovered the issue
From https://doi.org/10.6092/issn.1971-8853/8350 which redirects to https://sociologica.unibo.it/article/view/8350, I can click on the pdf icon to get to https://sociologica.unibo.it/article/view/8350/8272 which I can download locally and then reopen in Chrome to annotate with the Hypothes.is client.
I was able to make an original public annotation: https://hypothes.is/a/Nysv1HyTEeyaC2cnv3ZCPQ
Having subscribed to my public individual user feed, this annotation (via the annotation permalink and not via the original document) was found in Ton Zijlstra‘s RSS reader, and he was able to reply to it: https://hypothes.is/a/p3uUBJc8EeyuRmfRyGEGfQ.
Oddly the URL https://sociologica.unibo.it/article/view/8350/8272 when activated for Hypothes.is doesn’t show any of the annotations though I would suspect that the .pdf fingerprint should match that of the downloaded and annotated version. Alternately visiting https://uni-bielefeld.de/soz/luhmann-archiv/pdf/jschmidt_niklas-luhmanns-card-index_-sociologica_2018_12-1.pdf shows 51 annotations in the Chrome extension, though none of them are visible and the .pdf file doesn’t load on the page which returns a 404. Ton Zijlstra, having none of these URLs would otherwise not have been able to find or reply to annotations I’ve made other than having the original pointer via his RSS feed.
This last part non-withstanding, after making his reply to my annotation (directly at https://hypothes.is/a/Nysv1HyTEeyaC2cnv3ZCPQ), Ton Zijlstra is now no longer able to find his original annotation in the https://hypothes.is/search online interface. It’s as if it’s completely disappeared as the main web search interface is unable to find it via username and/or tags and (likely by design) the main public thread only shows top level annotations and not replies.
I’ve tried some similar experiments on my own replies to annotations. I’m unable to search my own annotations (via https://hypothes.is/users/chrisaldrich) or use either a user-based or tag-based search to find those same annotations after they were made, thus they’re essentially lost to me and others unless I can find the original document and trace my way back to them. These replies are obviously available via feeds (RSS/ATOM) and the API (using the urn:x-pdf:471902ab75f5683c53518d14f95f0dfe key), but they are essentially lost to the vast number of users who won’t have recourse to these methods.
Similarly searching Ton Zijlstra’s user name: https://hypothes.is/users/tonz, one will see no public annotations despite his public reply to a public annotation. The reply can be found at https://hypothes.is/stream.atom?user=tonz and via API calls.
After having made a reply to an annotation (with or without tags), one should expect to be able to search their own annotations or specific tags and find those public replies to annotations again.
Whether or not the main web stream (https://hypothes.is/search) filters out replies, they should still be able to be found via subsequent direct search.
Searching for one’s previous replies, via user, tag, or otherwise doesn’t find them, though they certainly exist and are findable in feeds and API.
Related, possibly helpful for the above
- Document fingerprint in the example above: urn:x-pdf:471902ab75f5683c53518d14f95f0dfe
- docdrop version of the same fingerprinted pdf: https://docdrop.org/pdf/Niklas-Luhmanns-Card-Index_-Th—Schmidt-Johannes-F.K_-0rcv9.pdf/
- Example extant annotation that doesn’t resolve on https://uni-bielefeld.de/soz/luhmann-archiv/pdf/jschmidt_niklas-luhmanns-card-index_-sociologica_2018_12-1.pdf#annotations_HzcGQHyFEeyQAqNHDXjO3Q
I’ve tried on other platforms and browsers and platforms with similar results, but I’m using Windows 10 and see the same behavior in both Chrome (Version 98.0.4758.102 (Official Build) (64-bit)) and Firefox (97.0.1 (64-bit)).
cc: Ian O’Byrne, Remi Kalir
@chrisaldrich[‘s] hypothesis feed is valuable to me.
Try it at https://annotations.lindylearn.io/page/?url=https://www.theatlantic.com/magazine/archive/1945/07/as-we-may-think/303881/