Algorithms are not just used by computers
In the media recently I have become aware of a peculiar change – the increasing reference to algorithms. The term algorithm is quite often used with a sense of awe or fear or both, and mainly with reference to Artificial Intelligence. The worrying thing is that it is beginning to seem as if algorithms only exist in computers and even that there are only AI algorithms. But humans have used algorithms for centuries, if not millennia.
Consider how, if we want to make a meal that we have not made before, we follow a recipe. When we do, we are in effect following an algorithm. If we want to construct a piece of flatpack furniture we are provided with instructions to follow – also a sort of algorithm. Want to play a tune that we have not played before? We follow a score – another kind of algorithm. An author wants to share a story that she has imagined so she writes a story. A story is a particularly ingenious algorithm that will lead us to exercise our minds interpreting the language-encoded-in-words in such a way that we get the story, because a text is a way of leading the human brain to have particular thoughts that lead to conclusions and emotions intended by the storyteller. I could go on, but leave you to come up with more examples. My point is that as humans we use something like algorithms quite often and to good effect. It is no coincidence that the word is derived from the name of a man who described a method to enable human computers to achieve particular mathematical goals. Following ordered steps to achieve something is part of how we are able to achieve so much of what humans are able to achieve. But only one part. We pride ourselves that we are able to use our discretion when following these rules because we know that human algorithms are for the guidance of the wise and for the obedience of fools, don’t we? We know we are following algorithms. We know when we should and when we shouldn’t and how and why they work, don’t we?
Do we understand the rules?
Well actually possibly not. I can make a mean white sauce but I do not know exactly why when I make a roux of flour and butter then add milk it does what it does; whether it works or doesn’t work depends on how well I execute the algorithm. If I play or sing along to a tune, I do not know why it moves me emotionally, making me feel joy or sadness the way it does, I just know that it does. So often we just follow instructions or training that we have been given or have learnt, trusting that it will all work out as we hope. But of those of us who have tried how many of us have not got a lumpy white sauce now and again?
Computing technology is just making the benefits and threats of this particular method of doing things more apparent. It is troubling that a specific technology may be blamed for problems associated with the use of algorithms instead of the concepts that have been embodied in them. Computers work the way they do because they are based upon just one sort of human thought – using algorithms of one sort or another. There are alternative ways of thinking and working and we need to remember and use them.
What should be understood is that although a sense of awe and possibly fear about what can be achieved with computer technology might be appropriate and no doubt needs to be dealt with, we also need to realise that the way humans make use of algorithms, in whatever context can come with problems too.
In short we need to be sure of what should be feared, what should be admired and why. Algorithms enable humans to do awesome things but also to do things that are aweful.
Humans are not digital computers
It is possible that some readers if they have got this far are fuming at the inaccuracy of my comparison of techniques used by humans to govern their activities with algorithms in programs used to control digital computers, because digital computers and people are different. Before continuing my argument I would like to try to calm any possible indignation. You will already know that there were computers before mechanical analog or digital computers – people who computed were called computers. This was alluded to above when I referred to the etymology of the word algorithm, so I shouldn’t need to remind you what the man whose name has been mapped into a mind-boggling range of ways of achieving goals in digital computing was aiming to achieve when he recorded instructions on how to calculate.
Enough of al-Khwarizmi. The purpose of my post is not to suggest that humans are computers, though they can sometimes compute. I want to draw attention to the idea that the use of algorithms in computers is founded on the use of algorithms by humans. Computers are sometimes like humans because they were designed to apply human ways of thinking, including step by step processes. Human ways of thinking are however are not always applied in ways that have led to admirable or supportable consequences. We need therefore to be very careful about how we apply and interpret the uses of computer technology. Not because it is a technology but because of the human concepts physically embodied in it.
Methods appropriate for managing digital computers may not suit humans
Ironically a fear of unknown processes determining computer behaviour is similar to the fear of human psychological behaviour, a fear that leads some people to want to govern fellow humans in a way that treats them as if they were machines.
If you learn a bit of computer programming and a little about how the digital computer actually works and if you understand how algorithms work in computers, you realise that the methodologies applied to address the problems faced in writing computer algorithms are not necessarily the best way to address human issues, because the problems faced are caused by aspects of the technology that are different to humans. You don’t need to understand how computers work at a fully comprehensive level – it would take up too much of your time that would be better spent doing other stuff, like cooking or playing music, taking part in sports and games, caring for others, making works of art and so on – but if you just grasp the basics you get an insight into just how flawed it might be to apply techniques used to get computers to function usefully to the organisation and management of humans.
As well as learning a tiny amount about computers and how to program them it is also essential to think more thoroughly about how we choose to use what computing technologies have to offer. This includes exploring why as humans we are so prone to following rules and fitting in with cultural expectations and so unwilling to use new guidleines (or algorithms) rather than the ones we are used to. For example, why we will ignore evidence that contradicts what we already believe?
A question of rituals and art
Let’s take a break from algorithms and computers and consider human rituals. A ritual is a sequence of actions intended to achieve a particular end. I got into serious trouble at art college once by suggesting that my tutors were taking part in a ritual when they assessed my work. At the time I was exploring how performance art fitted into the history of human ritual practices but it seemed that my tutors had little interest in considering the way that I was approaching it. The thing that fascinated me at the time was how particular courses of action in everyday life, when repeated often enough, could become like a ritual performance that might be observed for its aesthetic qualities.
We seemed to think so little about the aesthetics of how and why we act the way we do. More to the point, I wondered, why do people value some repeated activities more than others, making some so important that they are willing to reprimand or even kill other people if they did not perform them and conform to particular rules when doing so? As I was studying art at the time my particular questions revolved around what constituted art. Why were some works considered better than others – and how did the activities of critical engagement contribute to our ability to evaluate the relative quality of art works?
I theorised that when we make art of any kind for people to experience we want their perceptions and thoughts about the work to build a kind of sculpture in the mind by changing or enhancing neural pathways in the brain. The medium chosen by the artist is used to cause experiences that will effectively sculpt the mind so that the spectator or audience is led to have a more or less specific aesthetic experience. My conjecture was that the material form of the work of art was only important in as much as it managed to achieve the necessary mental changes. The challenge was to work out what changes, in what medium were necessary to cause the desired effect.
A glitch in this algorithm
Oh dear, I fear I may be drifting off topic. But I think it is a clue to the issue that concerns me and I shall return to it later. It can get confusing if banana elf scrimbongle doodlywibble dupdup noodlewarble.
See what I did then? I introduced a deliberate error in the code of this text-algorithm.
Interaction and participation lead to new questions
Let’s recapitulate. Humans can sometimes achieve things by following rules. They have therefore built machines so they can also achieve things by following rules. Following some rules can lead to mental changes that lead to aesthetic experiences.
A common goal in making art has been to create illusions. Regularly people willingly suspend their disbelief , as Coleridge put it, and go along with the artist, in his case the poet. People participate in realising a work by accepting the proposal that something is different than it might otherwise seem. They have to interpret the work presented to them.
To engage people knowingly in the realisation of my art I started to make pieces that were intended to be manipulated and used to make compositions, with the idea that it might lead people to become part of spontanious ritual-like art performances. Later I began to use computer graphics and felt that interactive computer-based pieces would be much more likely to engage people in participatory works. It also became possible for me to develop work derived from an interest in the aesthetics of social interaction that had begun when I learnt about body language from the work of Desmond Morris. Among my first thoughts when I had the chance to try computer programming was that as the logic of computers and programming had been invented by humans, by learning how to program I would learn more about the logic followed by humans and perhaps about how we decide what is art or not. As a result of learning how to program and trying to make art using systems that are based upon viewing the patterns of social behaviour of animals and humans as well as being a participant in the recent changes that have happened in the management of the rituals of academic life, I have come to develop new questions and make connections between the relationship between human and machine algorithms. I have been concerned, like so many others, about the way that computer technology is having a profound effect on our society at large. My opinion is that algorithms appropriate for machines are being applied to humans. I also believe that this is being misinterpreted as a new phenomenon. I have a hunch that it has roots in some pre-historic and possibly fundamental human behaviours.
My hypothesis is that basic corraling and animal management techniques led to a form of data processing, where the data were cattle and the cells were animal folds. I will discuss this a bit more in my next port. But because it has been used more recently to embody just some human thinking methodologies, those related to calculation and data processing, it leads to the techniques used to get programs (software or apps – call it what you will) to work gaining an importance it makes them seem more significant than others.
The fact that algorithms exist in digital computers is attracting attention it should with luck lead to a deeper potentially more useful discussion about the way that humans have used algorithms in a broader sense for millenia.