Once I was talking with a friend about an upcoming career decision. I said I wanted to do something “valuable” and he said that he recognises value when he sees it. Understanding values has never been easy for me, so I was intrigued by his sentence. I started asking myself some questions. Why doing something valuable felt much harder to me than to him? How did our approach to understanding what is valuable differed? Is determining value an objectively hard problem or not?
To start, I reviewed and reflected on some common ways in which we attempt to understand what we and others value.
How many times when you asked a friend for help on a decision they answered you “it depends on what you want”? How many times after that question you started describing that wanted to start that business or how you would feel incredibly good if only you found a partner that was right for you?
These are examples of explicitly describing what we value in words. It is the most common way in which we find out what we value: by directly answering “what do you want/value?”. Sadly, determining value in this way is often inaccurate. Very often, we say we want things that we don’t really want, for many reasons. Other times, we want things that we can’t admit we want. Derek Sivers gives a great example of this from his personal life. This is also why surveys and interviews are often a poor way to understand what someone wants.
Similarly, when people engage in life coaching, they usually start by finding out what they value. There are different ways to do this: shortlisting people we look up to and working out what they represent for us, go through a list of value words and pick what resonates with us, comparing values we like and selecting which ones we would choose even over another value we care about and many other methods.
I tried several of them. The result is a list of abstract ideas I felt somewhat emotionally attached to. However, I rarely refer back to it. It is rare that I have to decide between an option that closely matches my values and one that doesn’t. More often, I have to decide between options that contain some mixed elements of most of my values.
This is true for companies as well. Startups are encouraged to clarify their purpose and values early on to establish a strong culture. At large corporations, values often take up a large portion of company handbooks and About Us website sections. Yet, many perceive that these values as meaningless and disconnected from our daily reality.
Overall, trying to figure out what is valuable by what we and others say and write seems often too high-level to be practical, inaccurate and quickly out of date. On the positive side, spoken and written values can sometimes act as signposts we can check our actions against in the future.
Many think that if people are willing to pay for something then it is valuable, and that’s it. In a way, this is an indisputable fact. If I am willing to pay 5$ for a sandwich, it is a fact that at the present moment I valued owning this sandwich 5$. At first sight, money seems more reliable than words as a proxy for value.
The whole point of money is to enable more exchange. It enables people to trade different things that they value. Thus, money is a way to understand and determine the value of things between parties. However, money does not help us much with understanding what we internally value. Since we cannot really exchange things with ourselves, there are many facets of what we value that money cannot capture.
One could argue that we can use the way we decide to spend money as a way to understand what we value, and that seems to be pointing in the right direction. However, the problem here is that there are many important decisions in life that don’t involve spending money. Why did I go out for dinner with one person instead of another, even if I would have spent the same 100$ dollars with either?
If how we spend our money can tell us something about what we value, then by looking at how we spend all our resources (money, time, effort, etc) we can understand what we value. That is, we can take actions as the proxy to understand what we value. At first sight, it seems like we are making progress. From words to money to actions, each of these seems to have a greater degree of accuracy in determining what we value.
However, the main problem of using actions as a way to understand what we value is that they don’t generalise. What I mean is that observing an action tells us hardly anything more than the person that took it valued that specific action in that specific context. We cannot infer that because I preferred to go out for dinner with Julia instead of Laura, I will make the same choice tomorrow. We cannot infer that because I decided to be honest today, I will be honest tomorrow too.
This is because of many reasons. First, the action we observed could have been a mistake (eg I wanted to go to dinner with Laura but confused phone contacts). Second, humans are often irrational, and while we may accept it, there is no clear way to describe and model irrational values. Third, humans learn from their actions. So whenever we take one, we may change our values as a result of the actions we took.
Self-observation and reflection
If we can use neither words nor actions (including spending money) as a way to understand what we value, is there any other way? Or are we doomed not to understand what we value?
In principle, it seems unreasonable to me to accept that we cannot understand what we value. We experience being alive, we do things and make decisions. As much as we are irrational, behind those decisions there are mechanisms that assign greater importance to one option at the expense of the others. Our lives seem too orderly and structured for every decision to just be random.
Furthermore, we are consistently exposed to those mechanisms. We witness ourselves living all along. Whatever we do, we are there to observe. Hence, the best strategy I have found so far to understand what I value is to observe my experience and to reflect upon it.
This requires bringing it all together. Observing what we say we value. Observing the actions we take. Noting how our feelings and internal monologue relate to words and actions. We can use words as a way to summarise what we understood about what we value so far. We can observe actions to notice inconsistencies and inaccuracies in our current model. We will never be perfectly accurate, but we can keep iterating. In a way, it may be an endless journey. We will never be done with understanding what we value, yet we cannot stop.
Clearly, we can be misled by our internal biases and misunderstand what we value. But spotting and reflecting on the inconsistency of our actions is part of self-reflection, so we are not completely screwed up by this, and we can gradually overcome it with careful observation over time.
The pains of a non-answer
From observing and reflecting on ourselves, we already understood a lot about what we value. We understood that most of what we value boils down to what we have associated pain and pleasure with (consequentialism and utilitarianism). Yet, we also learnt that there are exceptions to this. There are things that we intrinsically value in their own right beyond pain and pleasure (categorical morality). An example is when we choose to love someone unconditionally not as a mean to pleasure, but as an end in itself.
In hindsight, I think I struggled with understanding what is valuable because I wanted a hard answer to it. I wanted an algorithm, a heuristic, a formula that would once and for all help me to have an easy time at decision making. On the other hand, perhaps my friend had already accepted that there is no such thing. By recognising value when he sees it, he implied that he is confident that by continuously observing and reflecting on his own experiences, behaviour and decisions, he could engage with things and understand if they were valuable to him or not.
Understanding value is a hard problem. However, there are two ways of tackling this hardness. On one hand, we can design more complex and sophisticated heuristics to determine value. On the other hand, we can accept to bear the pain of endless self-observation, reflection and update. Once we realise that there may never be a good enough heuristic, we realise that these two options are the same thing.
Separating personal and social values
Understanding value through self-observation and reflection has a major limitation. We can only use this for ourselves. We cannot observe other people’s internal monologue, feelings and reflections. In their behaviours, we only observe the external manifestation of their internal reality. So, does that imply we cannot determine what others value?
I see this question in another way. Perhaps understanding what other value is not our responsibility, but theirs. Aren’t we pushing ourselves beyond ourselves by taking on the hurdle to understand what others value? Aren’t we going on the other side of the net?
Can we live in a world where we accept we cannot understand what others value? Yes and no. On a personal level, we can probably bear the fact that we cannot fully understand others, that there will always be a gap and a net not to cross. This is part of respecting others. However, on a social level, we still need to make common decisions, plans and taking coordinated actions. However, this may be a different problem. It is the challenge of designing systems that can aggregate personal values effectively. We already have some: markets, voting systems and more.
We can separate these two problems. We can leave the problem of understanding internal value to each of us, to the growth and development of our consciousness. And we can take on the problem of understanding and aggregating social value without the need of understanding each individual value system, but rather by letting each one of us express their own values and engage in a way to coordinate them.
The codification of values
If we accept that we cannot fully understand the personal values of a being from outside of it, but simply engage with it expressing them, we stumble upon a challenging implication.
As technology becomes more and more powerful and Artificial Intelligence demonstrated unprecedented information processing and problem-solving abilities, we already experience and start to envision times where technology needs to act on behalf of someone. This requires a solid understanding of the values of the principal we acting on behalf.
In this scenario, assuming that we can extract values from people and beings in a hard and formulaic way will naturally lead to problems and limitations to the applicability of Artificial intelligence and automation technologies.
Accordingly, we will need a way to interface with the complexity that AIs will be able to model but that will not fit into our less powerful physical brains. We will need to find ways to discourse, engage and communicate back our self-observations and reflections into the algorithms. While this may sound inefficient to many, given the physical limitations of our brain, there may be no way around it.
However, it is important to note that some believe that AIs will be ever so powerful to understand our values from outside of ourselves. That they will recommend things to us and we will don’t understand why, but we would follow them because we would have accepted that they know best. I don’t know if this is true and I don’t push myself to make a forecast. If it is, then my reflections here may be irrelevant.
Whether machines will somehow evolve beyond mere tools into full beings, remains the most important mystery of our era. My sci-fi speculation is that it may require us merging with them to equip the union with powerful information processing capabilities and the drive to make meaning that comes from being alive.