Philosophy for architects: where did all those reasons come from?

Let’s do some philosophy!

When people notice that I have a PhD in philosophy, specialising in ethics, they often ask me how I apply it in my day job. Up until a few years back, the honest answer was that I didn’t: I tried to behave ethically, like everybody else, but that was just normal behaviour, not an attempt to do philosophy at work. That has changed following the growth of commercially useful artifical intelligence: interest in AI ethics has brought philosophy into the workplace. I hope that it has shown that philosophy does not have to be an obscure, impenetrable discipline: at its best, the aim of philosophy is to answer hard and important questions that matter to everybody.

So, now that philosophy has come to the office, let’s see if we can put it to work. We can start by seeing if when we need to make a decision, but people just can’t agree with each other. (I realise that philosophy is not for everybody: if it’s not for you, skip to the TL;DR non-philosophical punchline at the end.)

The area of philosophy I am most interested in is a branch of ethics known as practical reason (see, it’s even got ‘practical’ in the name!). In this area we try to figure out why we have reasons for action. Why do something rather than do nothing? When can an action be considered irrational? What does it mean when we disagree over what we ought to do?

One particularly big question in this area is: where do our reasons for action come from? Do they have an ultimate foundation in human nature, or the nature of the universe, or even the nature of logic itself? That would be a reassuring conclusion: if we knew that there was some ultimate foundation for our reasons, then our job would be to discover the right reasons, supported by that foundation. If we disagree with each other, then we simply have to figure out which of us has found the right reasons and which has not (or whether we are both mistaken).

Those of us who are technology architects often behave as if we believe this conclusion. We start by asking questions about action (What product should we buy? What technology should we use? Should we adopt this architectural style or that architectural style? Should we organise our work in this way or that way?) and we proceed by trying to answer these questions through analysis (What features does the product have? Does this technology conform to our standards? What are the performance characteristics of this component? What is the efficiency of our organisational structure?).

We hope that this analysis will result in data, facts and conclusions which, when organised well and presented clearly, will yield clean, simple decision making and universal assent. Yet such an outcome is often elusive. Instead, we frequently find that analysis is not the end of the conversation, but the beginning of the conversation, and that the conversation is suddenly full of emotion: affection for existing systems; fear of the unknown; excitement about the new; impatience to achieve results; frustration at the speed of decision making; caution about going too fast.

As technology architects, we are (usually) rationalists and technocrats, and, confronted with such emotion, we are tempted to reach for the favourite tool in our toolbox . . . more analysis! And we are puzzled when this tool does not always work.

I believe that more analysis and more reasoning do not always work because, whether we realise it or not, we are making a philosophical mistake: we are operating under the belief that our reasons for action come from somewhere outside us, somewhere purely objective. I subcribe to a different philosophical position: that our reasons for action are dependent on our non-rational motivations (emotions, desire, wishes, feelings and so on).

As with most philosophy, there is plenty of disagreement with this position: some would argue that it undermines the objectivity of reason, or that it leads to a sloppy relativism. We could even disagree about whether this topic even belongs in the domain of philosophy: perhaps it could be better addressed through psychology or neuroscience.

Let’s put all that disagreement to one side: we don’t have the years of time or gallons of ink necessary to explore it properly. Instead, let’s ask whether this insight can be of any use to us.

I think it can. We may not agree that all reasons stem from non-rational motivations, but I think that it is reasonable to accept that some reasons stem from non-rational motivations. I also think that it is reasonable to accept that some of those non-rational motivations should be respected. And I think that it is reasonable to expect that some of those respectable non-rational motivations will not be changed due to reason or analysis.

For example, if one stakeholder cares about customer experience more than anything else, this does not always have to be justified in terms of revenue or other secondary metrics: they may just care about customers more than anything else. If another stakeholder cares about the wellbeing of the team more than anything else, this does not always have to be justified in terms of productivity or other secondary metrics: they may just care about the team more than anything else.

It is important for those of us trying to lead people to decisions to recognise when we have reached these strongly rooted reasons: those reasons which cannot be justified in terms of anything else. There is no point in doing more analysis to resolve clashes between such reasons: there is no analysis will not change minds at this level.

The best thing to do in such situations is to make the underlying motivations explicit so that we can at least figure out what to do about them: this won’t be easy, but it is more likely to yield results than analysis proceeding on the basis of a mistake. If we are going to disagree about things that we are never going to agree on, then let’s at least know what we are disagreeing about.

TL;DR non-philosophical punchline: when trying to make decisions, no matter how good our analsysis, we will find sometimes people who disagree because they have different motivations at a deep level. More analysis is not the answer to that problem: we must understand and address those motivations.

It may seem unlikely that we will encounter such depth of feeling when making decisions about technology in a corporate setting. However, if you have ever tried to facilitate such a decision, you will know just how deep some feelings run. Furthermore, if we are going to ask people to bring their whole selves to work, we have to expect that those selves will include some things they care deeply about.

If you found this little bit of philosophy for technology architects interesting or useful, please let me know: I think that there are other questions worth exploring.

(This blog post was inspired by the essay ‘Internal and External Reasons’ by Bernard Williams, published in Moral Luck, Cambridge University Press, 1981.)