Our World is Getting a Little Weirder with Argument Mining
What are you doing when you are involved in that very human activity of using your common sense? How does your brain construct an argument in several seconds, minutes or hours before you feel compelled to speak your mind? It all makes sense inside your head whether you can explain it or not, and so that is usually the end of the discussion.
Welcome to the twenties — things are about to get weirder. As scientists race to strike the metaphorical mineral veins of the human mind, many are picking up their pickaxes and pointing them directly at human logic. If you’re angsty about the fact that your skills as a licensed driver may soon be automated away, you may curse out loud upon learning that your cherished argumentative skills may too. Humans may want to shout a bit louder to slow down progress of automating away the latter.
Argument mining is a technique that is rapidly growing in trendiness among forward-looking researchers and entrepreneurs in the tech industry. By building software to scan text and extract sentences that it identifies as evidence, such tools can demonstrably be trained to assemble arguments of their own. These tools may point out factual inaccuracies in fake news articles, assist students in rewording essays to be more convincing, or snobbishly cancel an absurd claim at the dinner table on one’s behalf. However, digging too deeply into the mines of our argumentative capacity might cause us to strike the stubborn crust of our imagination — and cracking it won’t be easy.
I’m a big fan of the arguments made by macro-historian and author Yuval Noah Harari. In my previous article about the importance of scientific accountability in the sci-fi community of this decade, I referenced a fundamental argument of his — fictions gave humans the cooperative and intersubjective edge required to control the planet and rise above all other species. Our stories until now have served us well as our minds have learned to adapt to our society’s growing complexity, enabling the construction of arguments of all kinds to suit our agendas. However, this advantage might soon become a burden.
We are now living through an age of scientific progress unlike any other before us. Our scientific discoveries are converging with the narratives born out of our imaginations, as Harari comments:
“As human fictions are translated into genetic and electronic codes, the intersubjective reality will swallow up the objective reality and biology will merge with history. In the 21st century fiction might thereby become the most potent force on earth, surpassing even wayward asteroids and natural selection.”
Will the automation of human arguments complicate future scientific and societal progress if we have yet to scrub out the stains of our as yet unrecognized fictions? While it may be trendy to pour our imaginative biases into the logic of machines that we build today, would future humans even notice if they get guided down a path of seemingly rational nonsense for years to come? Naturally it would not be nonsense to our descendants — it would take the form of a strong argument that was constructed by the far more logical machine.
If you can imagine a future in which human logic can’t keep up with the ever increasing intelligence of our logical machines, you can perhaps imagine a different future in which humans are allowed the freedom to revel in any degree of shallow logic that pleases them. Which would you prefer?
It feels counterproductive for argument mining to peak in trendiness at any point in time. Perhaps we should view it as a field that needs to grow gradually with our ever-developing reason, rather than one that needs to be hastily dug out for all its treasures before the nation next door beats us to it. Adopting this view means that our understanding of an argument gets to progress across the patient plains of time — through all the exciting discoveries that an imagination may present to our descendants long after we are gone.
