Cheapening the Written Word

The last two years have seen excessive hype over the text generation functions that large language models facilitate, which doesn’t need to be remarked on here more than it already has. But, I do think it’s important to note that applications like word processors and email have been transformative for the practice of law and other knowledge work over the last 40 years, and this can be considered an expected continuation of this long term trend.

These types of tools all reduce the friction involved in creation of documents and mean that written material can be produced more quickly and cheaply than before. If we make errors in our typing we don’t need to start again to make a fair copy, and we don’t need to pull out scissors and glue to copy and paste content within our documents. However, instead of resulting in a more efficient system, it appears this added facility has simply increased the volume of documents, which means that arguably the potential improvements and access these tools promise haven’t been achieved.

In an archival course I took years ago, I remember studying a court record from around 1900. It was a court record documenting the court process surrounding the criminal matter for someone who had been charged with stealing a bicycle. And as a record it was perfect: it was one page long, with fields where staff could add notes with typewriters or by hand, and the final verdict was handwritten in pen at the top. The entire matter was visible on one side of a piece of paper. Old court reports are similarly brief. The details included about cases in Old England Reports often run only half a page. Meanwhile, a contemporary trend for court decisions to get longer has been documented.[1]

Since then, writing technology has progressed, and almost all professionals have access to computer-driven applications that allow quick production of documents that have been spell checked, and do not need to be re-typed if edits need to be made now. As technology has advanced, however, instead of making processes more efficient, it has just allowed for long documents to be created cheaper than they were previously. The decrease in the cost of making documents longer has led to increased complexity in process and communication, as there is no longer pressure to put in the work required to be concise.

In the legal system, this means that the efficiency gains that could have been had have instead been directed at making legal systems more difficult to navigate. This effect can be illustrated by the need for systems like e-discovery. When I hear about e-discovery I imagine all the hours of labour that went into creating those written documents, and of course, now content doesn’t require human intervention at all.

Writing is a particular type of work, and much of the work in the legal sector is comprised of writing, but accounting is an interesting comparator. Significant work and expertise used to be required to maintain simple bookkeeping functions, as typing used to require more skill than it does now with spell check and sensitive keyboards. However, the invention of spreadsheets devalued much of this work and made it much easier to track financial position. The reduction in requirements to maintain manual calculations allowed accountants to create significantly more value through approaches like financial forecasting, which allowed the view of the profession to expand.

Law has not been successful at achieving similar increases in value, because the base is about human relationships, and it has been resistant to change of this kind. At the centres of contracts are not documents, but agreements between people which mean they both get something. Complex language can also be used as a weapon, which reduces some of the benefits of simplicity.

The next great transformative technologies that can be leveraged to make language creation easier or faster or cheaper based on large language models may have significant impacts on making the practice of law more efficient. It may also be used to further the legal professions’ impulse toward completism and simply lead to everything being longer and more complicated. In other words, they may be used as blunt objects to increase the volume of words, without taking the time to value the virtue of brevity.

An internet search for the source of the quote apologizing for writing a long letter because the author didn’t have time to write a short one, shows that this has been attributed to many writers going back to Roman times. And it appears to have originated with Cicero. Now, we have tools that will generate text even more easily than word processors and dictation, but the mental discipline of saying what needs to be communicated and cutting out the extraneous will continue to require effort.

[1] Examples: Beauchamp-Tremblay, Xavier, and Antoine Dusséaux. “Not Your Grandparents’ Civil Law: Decisions Are Getting Longer. Why and What Does It Mean in France and Québec?” Slaw (blog), June 20, 2019. https://www.slaw.ca/2019/06/20/not-your-grandparents-civil-law-decisions-are-getting-longer-why-and-what-does-it-mean-in-france-and-quebec/. SCOTUSblog. “Lengthier Opinions and Shrinking Cohesion: Indications for the Future of the Supreme Court,” July 28, 2022. https://www.scotusblog.com/2022/07/lengthier-opinions-and-shrinking-cohesion-indications-for-the-future-of-the-supreme-court/.