Jun 26, 2026 · 30 min · 14 segments
The Bentham Project oversees the publication of The Collected Works of renowned philosopher and reformer Jeremy Bentham. In this episode, Professor Judy Stephenson hears more about the project from…
Judy StephensonHostAnd that ranges from short pamphlets of 12 to 16 pages, right up through the rationale of judicial evidence, which is 2,500 pages long.
Starting with manuscripts, the first thing we have to do is get an accurate transcript of them and there's a whole series of processes which we have to go through to take those manuscripts and turn them into a coherent text.
Just reading Bentham's art writing is difficult in itself and with crossings, interlineations, notes in the margin.
And you don't start at page one and go to page 200 and oh, there's your text, because the stuff's all mixed up and Bentham would write and rewrite.
how Bentham himself was constructing that text at each major stage in its development.
So I'm a historian and I work with archival sources a lot and people say to me, oh, but AI does this for you now.
So can machine learning or artificial intelligence help with the transcribing and the gathering before the analysis? Or what contribution can it make to this sort of
work? We, as the Bender project, were very much involved in the Transcribers project, which applied artificial intelligence to reading unwritten text.
But the people we work with, the computer scientists and other technicians, took Bentham's manuscripts as particularly interesting because they were challenging in various ways.
Yeah, this is our experience on things like customs records, on other archival sources.
And it makes inhuman assumptions about the relationships between different things on the page and stuff as well.
I mean, the way we operate is we have Transcribe Bentham, which is a scholarly crowdsourcing initiative which has now been going 15 years, probably one of the longest running digital humanities projects.
And that ranges from short pamphlets of 12 to 16 pages, right up through the rationale of judicial evidence, which is 2,500 pages long.
Starting with manuscripts, the first thing we have to do is get an accurate transcript of them and there's a whole series of processes which we have to go through to take those manuscripts and turn them into a coherent text.
Just reading Bentham's art writing is difficult in itself and with crossings, interlineations, notes in the margin.
And you don't start at page one and go to page 200 and oh, there's your text, because the stuff's all mixed up and Bentham would write and rewrite.
how Bentham himself was constructing that text at each major stage in its development.
So I'm a historian and I work with archival sources a lot and people say to me, oh, but AI does this for you now.
So can machine learning or artificial intelligence help with the transcribing and the gathering before the analysis? Or what contribution can it make to this sort of
work? We, as the Bender project, were very much involved in the Transcribers project, which applied artificial intelligence to reading unwritten text.
But the people we work with, the computer scientists and other technicians, took Bentham's manuscripts as particularly interesting because they were challenging in various ways.
Yeah, this is our experience on things like customs records, on other archival sources.
And it makes inhuman assumptions about the relationships between different things on the page and stuff as well.
I mean, the way we operate is we have Transcribe Bentham, which is a scholarly crowdsourcing initiative which has now been going 15 years, probably one of the longest running digital humanities projects.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.