Performative reasoning
Photo credit: Liam McGarry via Unsplash
How many of your reasons are comforting illusions? How many of the decisions which your organisation makes are based on well-ordered reasoning, and how many are simply surrounded with the trappings of reasoning in order to make you feel better?
Organisations make a lot of decisions. What products to buy, what products to launch, who to hire, where to invest, which projects to support and which initiatives to cancel. These decisions are particularly apparent in the field of enterprise technology, where we make choices about how to design, build and operate systems, how to organise resources and how to adopt and integrate new capabilities. The need to tell machines precisely what to do seems to require precision in our own thinking.
Because these decisions seem important, we feel that we should take them seriously, and be seen to take them seriously. When we are making purchasing decisions, we construct elaborate scoring criteria, invite bids and conduct extensive evaluations. When we are planning investments, we build detailed business cases, evaluate ROI and risk factors, and construct portfolios of change. When we are running delivery programmes, we create dashboards, produce reports and run steering meetings, so that we can respond to circumstances and keep everything on track.
And yet, despite all of this apparent reasoning, we find ourselves with purchases we regret, investment portfolios full of duds, and programmes which career out of control.
Sometimes this is simply because decision making is hard, and it is inevitable that we will get things wrong. Even if all of our reasoning and analysis is thorough, diligent and sincere, we cannot know everything, and we will sometimes make mistakes.
However, sometimes it is because our reasoning isn’t reasoning at all. The purchasing process is designed to pick the product which the most senior person prefers. The business case is based on wishful thinking, magic numbers and mystification. The reporting regime for the delivery programme is intended to keep sponsors happy and avoid upset and drama.
This may sound cynical, but I suspect it is familiar to anyone who has worked inside a large organisation. Even if it is never said out loud, there is always the suspicion that overt mechanisms of reasoning, analysis and decision making are smokescreens: means of generating apparent credibility for decisions which have already been taken, or realities which no-one wants to face.
If we know this to be the case, why bother? All the trappings of corporate reasoning take time, effort and resources: the purchasing process may take longer than the implementation of the product; the construction of the investment portfolio may extend far into the next financial year; and the business of reporting progress can get in the way of making progress. Why not save time and just let the sponsor pick the product, let everybody pursue their favourite project, and admit that we have no idea when the delivery programme will finish?
I believe that it is because people want something from reasoning. It would be best, of course, if they wanted to have good reasons: to be right more often that they are wrong. But it is also because they want reassurance (if we have done lots of analysis, perhaps we have found a good answer), or because they want to defer accountability (if we have followed the process, no-one can blame us for being wrong), or because they are uncomfortable in trusting expertise (a scoring sheet appears more reliable than one person’s opinion).
These desires, for reassurance, blamelessness and reduced reliance on experts may explain an aspect of LLM usage which I must admit that I struggle with. Lots of people, including senior corporate decision makers, appear to be using LLMs as ‘thought partners’, or even as thought substitutes. Sometimes this is covert (an email or an instruction which appears original but looks increasingly AI-like on closer inspection) and sometimes it is overt (an appeal to the supposed authority of output from AI). I struggle with this use of AI, because I value the process of thinking things through, of grappling with difficult concepts and making hard choices, and I find it hard to trust slick, glib, homogenised output. But perhaps this phenomenon is simply a replacement of mechanisms that looked like reasoning (analysis, scoring and evaluation) with a new mechanism that looks like reasoning (use of a model). In both cases, the aim may not be to reason, but to give a performance that looks like reasoning.
If you suspect that you are indulging in performative reasoning in your organisation (and I know that I have done so many times in my career), whether facilitated by AI or by old-fashioned corporate processes, there is good news and bad news. The good news is that the remedy for performative reasoning is readily apparent: do actual reasoning. This does not mean more spreadsheets, more scoring and more analysis: it might manifest as staring out of the window, scribbling on sheets of paper, or figuring out which experts to trust. The bad news is that actual reasoning is hard work, it might take you to places of doubt and uncertainty, and you carry the accountability for your thoughts and decisions. But I believe that the work is worth doing, and that it is better to be thoughtfully uncertain than performatively confident.
