Seeing an institution’s memory as process, not thing
Kusaladana — a discussion paper for the Science and Buddhist Practice retreat, Adhisthana, October 2026
1. The situation
I have recently taken on IT support for an order office. The first job is ordinary: move decades of accumulated files from Dropbox to Google Workspace. But sitting inside that ordinary job is a problem every Triratna institution will recognise.
Nobody knows what is in the archive. Folders were made by people who have moved on, following logics no one remembers. Information sits in silos formed around roles that no longer exist. The real map of the institution’s memory lives in one place only: the head of the administrator who answers the questions. When that person eventually leaves, the map goes with them.
The usual response is a tidy-up: sort the folders, impose categories, delete the junk. This paper suggests that response rests on a mistake.
2. The wrong frame
A tidy-up treats the archive as a thing — a warehouse of objects to be arranged on better shelves. But to arrange objects you must already know what they are and what they are for. With a legacy archive, that is precisely what nobody knows. Imposing a filing scheme on material we don’t understand doesn’t organise the archive; it just replaces one forgotten logic with another that will be forgotten in turn.
We are reifying the archive. And reification, we might notice, is a habit the Dharma has something to say about.
3. The turn
Try a different view. An office is not a store of things; it is a process. Requests arrive — an ordination enquiry, a question about the 2019 accounts, a safeguarding concern. Responses flow back out. The office is characterised not by what it contains but by how it responds.
Under this view every document changes meaning. Each file is the sediment left behind by a past transaction: someone asked, someone answered, and this document is what precipitated out. The silos are not storage failures — they are channels worn smooth by repeated flow, like water finding the same path down a hillside. They record where the recurrent questions have been.
This is not a metaphor imported from physics onto the Dharma. It is the same seeing arrived at from two directions. The archive does not exist as a self-standing thing; it dependently arises from the pattern of requests and responses that made it. Documents conditioned by questions, questions conditioned by need. Anatta, applied to an institution.
My own formation is in experimental physics, and the method here is a physicist’s: don’t characterise a black box by prising it open and cataloguing the parts — characterise it by its response. The proposal is to extend that method into meaning itself.
4. What this makes possible
Modern AI gives us, for the first time, an instrument that can read meaning at scale. Concretely:
A language model reads each document — or each email, each message — and writes a short free-form description of what it is: what transaction it records, what it concerns, what tone it carries. No fixed categories are imposed; the descriptions are allowed to be inconsistent, because meaning-similarity, not matching labels, does the organising. The descriptions are then embedded as vectors in a meaning-space and clustered.
What emerges is a map nobody designed. Dense regions where the institution’s life has flowed. Dry riverbeds — material no question has touched in years. And, most interestingly, gaps: recurrent questions for which no document holds the answer, because the answer has always lived in someone’s head. The gaps are a map of the institution’s tacit knowledge — the thing we most need to see before it walks out of the door.
As a first trial, I am running this on my own email archive — a quarter century of one person’s transactions — processed entirely on local hardware, with only aggregate patterns ever leaving the machine. Early results can be shown in conversation.
5. Open questions
I don’t bring a finished system. I bring questions I cannot answer alone, and this retreat holds an unusual spread of the minds they need:
- For the psychologists: an institution’s correspondence is something like its stream of consciousness. What would a clinical eye look for in its affective texture over decades? What would institutional health, or distress, look like in such a record?
- For the philosophers: what exactly is the ontological status of the “map” this produces? Is a cluster in meaning-space a discovery or a construction — and does the Madhyamaka answer differ from the analytic one?
- For the technologists: what does responsible custody of such an instrument look like inside a small religious charity — access, oversight, and the design of a system that must survive its builder losing interest?
- For all of us: Sangharakshita warned against “scientific Buddhism” while urging deep engagement with the cultural forms of our time. Does turning scientific method toward meaning — rather than reducing meaning to mechanism — fall on the right side of that line?
If any of this catches, I would like the conversation to outlast the retreat. The Lunar Men met monthly and wrote letters in between; their objects of study were half-finished things they could get one another’s hands into. This is offered as such an object.
Kusaladana is a retired physicist (semiconductor device physics, GEC Hirst Research Centre) and a member of the Triratna Buddhist Order. Background work on meaning-space and contemplative AI is at kusaladana.co.uk.