An exquisite irony
Johannes Trithemius was the Benedictine Abbot of Sponheim from 1483 to 1506. He was a polymath, a builder of libraries, and a deeply controversial character. He dabbled in the occult and was accused of inserting what we might now call “hallucinations” into his histories.
In 1492 he wrote a short book, De Laude Scriptorum (Eng: In Praise of Scribes). Movable-type printing having arrived a few decades earlier, his book was published in printed form in 1494. With exquisite irony, it defends the art of scribing at exactly the time that Gutenberg’s press was collapsing demand for handwritten texts and rendering scribes obsolete.
The book is sometimes held to be a medieval equivalent of a Luddite rant. Trithemius would have witnessed the economic impact on his own abbey. But his book is more than that. He appreciated the power of print to disseminate and democratise knowledge. His arguments tried to balance that against the organisational, professional, moral, and spiritual logic embedded in the older system of production.
Institutional memory
His first claim was about institutional memory. Without the scribe, he wrote, scholarly wisdom would never reach posterity. Writing gave “lasting value to passing things“. Without the written word, he warned, “justice [would be] lost, the law confused and the Gospel fallen into oblivion“. Printing, however, had the opposite effect to what he foresaw. It radically democratised written knowledge and drew into literacy many who had never read before. For lawyers, though, his raw claim is intuitive. Law is a system of memory and legal writing depends on faithful transmission. AI can retrieve, summarise, and generate legal language at a scale no human team can match. But is it reliable? Perhaps the question that Trithemius forces on us in this context today is a different one: what happens to legal memory when production is detached from disciplined professional formation? Following that, how can legal memory be safeguarded as AI continues to advance both in scale and complexity?
Durability
Trithemius’s second claim was about durability. Paper, he predicted, would quickly disappear. Parchment endured. He was only partly right. Many printed books have endured and the oldest are now treasured. This historical record shows, though, how often what was held to be true at one time has been shown to be false and has been supplanted by new, better knowledge. Clearly, clinging to old knowledge is risky in the absence of contemporary proof of its continued truth. This raw claim also translates well to the legal, digital world. The modern contrast is paper versus digital, and digital records can be deleted, altered, or lost with an ease that should trouble any custodian of legal memory.
A deep anxiety
Trithemius’s deeper anxiety was that the change in production technology would weaken the discipline that made texts worth preserving. Scribal labour added value, he thought, because it demanded attention, care, and inward absorption. “Every word we write is imprinted more forcefully on our minds,” he wrote, “since we have to take our time while writing and reading.”
This is perhaps his most directly modern, important point. Lawyers learn through friction. This friction comes through drafting, checking, revising, comparing, marking up, and being corrected. The process is often inefficient, but not all friction is waste. Some is cognitive apprenticeship. Some is ethical formation. Some is how professional judgement becomes embodied, and how the best answer to hard problems often emerges reluctantly and slowly – the slow thinking that Daniel Kahneman describes.
In this, Trithemius speaks most directly to firms wrestling with AI. Lawyers can see that AI may widen access to justice, improve consistency, cut cost, accelerate research, and help solve long-standing problems in service delivery. They welcome that. But they also worry, with very good reason, that if AI absorbs too much of the work through which junior lawyers become senior lawyers, the profession might lose the human capacity that gives the most difficult work (especially) its authority.
The need for disciplined human work
To reiterate, Trithemius’s book does far more than complain about the loss of his abbey’s revenue stream. He sought to defend copying as a practice that not only sustained institutions but also carried intrinsic societal worth. It fitted the monastery, he argued, because it aligned labour, discipline, learning, service, and identity. His most trenchant passage anticipates the obvious criticism. Why copy at all, he asks, when printing exists and “has brought to light so many important books“? His answer was not to reject printing. It was to deny that abundance removes the need for disciplined human work.
This parallel with law is direct. Legal work also combines knowledge production and formation. Associates produce documents and learn judgement. Partners supervise matters and transmit learning and standards. The firm sells advice and institutionalises its culture, risk appetite, and ways of producing that generally. AI need not destroy that, but it turns it into a design choice rather than an inherited feature of a now-outdated business model.
The scribe, Trithemius wrote, “is by no means defeated by the printer“. What he could not have foreseen, though, was just how far the democratisation of knowledge would run. The literacy of a whole society, rather than a small elite, produced benefits that dwarf what was lost. The Renaissance blossomed. Economic growth boomed. Entire new industries were formed. New forms of government emerged. It is hard to imagine how different the world would be today had printing never been invented.
Likewise, the lawyer is not today defeated by AI, and democratisation of the law will likely yield benefits and also challenges that are impossible to perceive ex ante. The essential task is to identify what carries the purpose that law serves in society. For instance: judgement, service to clients, trust, accuracy, confidentiality, ethical restraint, institutional memory, and the formation of successors. None of these survive automatically. It has to be deliberately configured into AI-enabled practice and perhaps even radically new models of jurisprudence that seem poised to be born. To add that all this will also affect the business of law – the business and operating models of law firms – is so obvious as to be trite.
A production technology is never merely that
Trithemius was wrong if he thought scribal culture could survive print intact. It could not, and scribes did not. But he was right in that a production technology is never merely that. It reorganises authority, memory, training, economics, and identity. That is why De Laude Scriptorum still matters. It is an early theory of professional displacement. At the organisational level, it warns that when a technology disrupts as deeply and as fast as AI is disrupting legal and other professional services, firms must be as quick to decide what to jettison as to redesign their business and operating models from first principles, so that what matters most is not lost in the passage.
This is the pivotal challenge. As AI takes over more of the visible production of legal work, as it will, how is professional authority to be generated, supervised, and transmitted? Who owns the judgement? How are junior lawyers trained once the apprenticeship tasks have been automated? How does a firm keep its distinctive knowledge durable rather than dissolved into generic machine prose, and remain more than a wrapper around tools available to everyone else?
The firms that find and implement the answers to these questions will likely be the industry leaders of the future and that, given the pace and scope of change, sooner rather than later.
Will yours be one of those?