I know so well that this is all there is
or the tyranny of the infinite knowledge graph and specialisation.
Let’s say you want to draw a knowledge graph (note: not the Google Knowledge Graph but something like it but in your brain). You want to put one together for tomatoes, I love tomatoes.
You might start with some basic truths (facts or attribute|value combos).
- Have the colour Red
- Have the types Cherry Tomato or Normal Tomato
- Grow on vines
- Are put in salads
We can draw that in a graph (not a numbers graph but a node/entity graph) below.
From the perspective of a general tomato consumer here in Australia, these are essentially truths. However, it does not take long before these truths breakdown as you dig deeper into the detail, even as an avid tomato fan I know that the first truth is incorrect, you may even go as far as saying it is a lie.
While most people would be happy to admit that tomatoes are not ‘always’ red, they may not always know that some can be purple, green, yellow and all are ripe! (depending on the type). That brings us to type, again we see the world initially with a few discrete types, but as you dig deeper and start purchasing tomatoes (not just consuming them), you understand that the choice depends on the season, location and budget you have to spend.
We can go deeper into each of our initial facts and our truths (which are really generalisations masked as truths) break down fast. Some tomatoes grow on trees , if you step outside of western culture you will see that tomatoes are not treated as a salad element but as a fruit in the same way one would eat an orange.
One way to look at this is generalisation, however I see this as a function of the knowledge graph, the closer you get to the core element of the knowledge graph the less nodes can connect and hence truths must be compresses.
As we go further into the knowledge tree of any area the somewhat confronting reality is that there is no end, and the truths at each level of knowledge continue to get broken down at the next.
To extend the tomato anecdote;
- If I want to start growing tomatoes at home, the types make much more difference, there are seasons, climate and soil/water issues to consider.
- If I ran a farm of tomatoes, this gets even more complex as I have to consider what can grow at scale, what market demand there is and again the types of vines/trees they grow on make a massive difference at scale.
- I could go even further and start a business to breed a custom type of tomato, this takes into consideration all the various breeds that are common, uncommon and even historic, what market there is in the farmers willing to grow and how to potentially tweak all the details that make up the truth of a tomato.
I think these levels loosely follow the path;
- Small scale producer
- Large scale producer
- Producer for producer
Each level obviously has multiple levels between, but we are only looking at a first round knowledge graph of how knowledge graphs could work :)
To streamline our levels we could look at a two axis matrix;
- Consumer — Creator
- Simple — Complex
At any point in the axis the truths appear to be set, but from the perspective of any other point, these are either complete unknowns (ie. The onlooker does not know what the Producers know) or un-truths of some scale (ie. the Connoisseur’s truths about a product make the Onlooker’s truths seem false).
Where Does this Leave Us?
Having a perspective that the truths in our place in the knowledge graph may be untruths at other points gives us the ability to be ok with being wrong and not knowing all things about all things. It also gives us the ability to empathise with others when they are working on a task that is much deeper in the knowledge graph than we work in day to day.
We can sum this up with simply ‘you don’t know what you don’t know’, it is equally empowering as it is sobering.
As we have an economy built around specialisation — where there is strength in each company and each individual doing a smaller set of things increasingly well — we must live with the reality that what we think we know at some point down the line is not considered truth.