Modes of Learning
How I Reached This Insight
I have two types of friends. One doesn’t believe anything you say unless it is backed by verifiable evidence or comes from a reputable source. The other listens to what you say, and then based on their model of the world so far, gives either their view on it, or gives counterarguments that might change my understanding of it. But in the end, the first friend stays at his position if I haven’t backed up what I’ve said with anything, and the second friend has updated his perspective with my perspective and I have updated my perspective with his perspective (and potential counterarguments).
I tried to get this point across to the first friend that there’s other ways of acquiring knowledge and socializing with the intent of sharing knowledge than just giving strong information.
I guess there are also parallels to be taken with people that hate smalltalk and people that use smalltalk as a probe to get the gist of the person before dwelling deeper into a topic.
Strong Information Learning
The world is a difficult thing to make sense of and as such we need to use tools that simplify this and make it more comfortable for us to interact with. For this we’ve developed accurate communications methods (like maths), formal logic and also words to describe our certainty level and ways of referencing sources - e.g. “I am almost certain that the name of the fungus Spongiforma squarepantsii refers to the cartoon SpongeBob Squarepants.”
So when you hear me say this phrase, you can simply update your world model simply and without issues mostly, because you trust Wikipedia as a source of truth and you also acknowledge the likelihood of this being true due to the sheer strength of the similarity of these words.
Strong information learning also happens for instance when you learn financial analysis methods from a book, test it out in practice and validate it yourself. You have little doubt of its applicability or at the very least, its relation to reality.
If you learn something by heart, for instance the periodic table and the corresponding atomic masses - this would also account for strong information learning as you need very little surrounding context to know all this. It might not be very useful without context, but you can still know this by heart and it might be useful in the future.
There are also people who use aspects of strong information learning only when it suits them in debates. For instance you say “I hate pineapple” to a pineapple lover and then they try to to make you into a liar because you didn’t specify your exact relationship with pineapples. So they reply “I’ve seen you buy and eat pineapple and say you enjoyed it!”. Yes, that might be also correct. When you said you “hate pineapple” you really meant that almost every time you encounter pineapple, it is canned pineapple and on pizza, which are the most frequent forms of pineapple. Still, you said something true, but not something that could be incorporated to the other’s knowledge base about you because they couldn’t take it at word value - they needed it to be more specific. This might often come across as them not trusting you.
Strong information learning can also be seen as building the world model bottom-up rather than top-down, but is fairly limited to what actually fits with the existing model. It’s like a jigsaw puzzle where the only way you can progress is by only putting the pieces together where you are certain that it’s correct.
Weak Information Learning
In contrast to strong information learning, weak information learning relies on a top-down buildup of a world model. This is more like building loosely related air castles of information and sometimes being able to connect them to each other. Or it’s like finding a green jigsaw puzzle piece, but you have no idea where exactly do they fit in the whole picture. The building blocks for these are various sources of information, both weak and strong, but most of all - plenty, so nothing will stay amiss and everything is incorporated that isn’t immediately obviously discardable.
With weak information learning you can incorporate gazillion sets of weak information interlaced together to create a decent model of reality. You can also entertain silly ideas honestly, openly splurt out opinions just to get validation or disapproval for the ideas, making the model more accurate in the process. The key here is the amount of information incorporated, rather than just using strong and verified information to update your knowledge base.
This mode of learning works best with most people in casual conversations as you cannot back up everything with sources all the time, or even barely remember where you got your ideas from. But it naturally does not work with people very well who are mainly strong information learners as you are being way too vague and seemingly just bullshitting if you have more uncommon beliefs.
Weak information learning is good for exploring and running around different topics until you get to a point where you’re more interested and perhaps want to know more deeply about. It is also a good way to explore in debates where you disagree with the other person, in order to pinpoint where you differ in your fundamental beliefs.
You can also have productive and useful conversations with people who are good faith actors who are speaking truth and want to explore topics together and share knowledge with each other, where you aren’t attacked on your potentially inaccurate statements (because you are honestly telling the truth and one potentially false point does not matter in a whole).
Hybrid Learning
There is of course a way to use both ways of learning and in all honestly, most people do. But I think they do this to a different degree at least within conversations. Mainly strong information learners are very annoyed with you if you come at him with the weak information learner strategy. And vice versa, with mainly weak information learners it might be too much of a hassle to talk to the strong information learners.
I think it is best if you adjust according to your conversation partner and do both according to the situation. If your conversation partner also knows both, you can explore topics with weak information style and once you get to a more specific topic that you want to dissect, you switch the mode.
But unfortunately most people are quite rigid in their ways, e.g. either they’re always 80% strong, 20% weak or e.g. 70% weak and 30% strong. I think it is valuable to understand that people are different in their communication and information exchange styles, so you can get along with them better and build more trusting, meaningful relationships.
Where Do We Go From Here
I don’t really know whether this theory has any larger context or applicability. It is just something I’ve noticed within the past few years and since I have to get along with people from both sides, I have had to adapt my style as well.
I tend to like being in weak information style most of the time as it feels more human, more emotionally connected to our beliefs, meaning you waste less time dwelling in topics you don’t care about as much (but can reason about).
Parallels
I recently found out (thanks to my cadillion) there are quite a lot of parallels out there in the world that might be related to the same phenomenon (but not necessarily).
Three Horizons
Even though this applies to mostly business, the similarity is evident.
This is from cadillion (edited for readability):
Three Horizons paradigm from The Alchemy of Growth talks about different stages of corporate development and the different abilities and metrics each stage requires of their leaders:
H3 is the explorer phase in which you try to find product market fit. It requires agility. flexibility, creativity, and willingness to fail before you find the approach that can be exploited. There is not always a lot of prior art to rely upon.
H1 is the exploiter phase in which you are simply trying not to kill the goose that laid the golden egg. Your market is well-defined, now you must simply relentlessly peel away inefficiencies and keep a large portion of the organization laser-focused on the developed approach.
H2 is somewhere between the two, where you have success, but it’s not enough to keep the lights on. Here you have some evidence that you are doing the right thing, but you are unsure if you should look for something more profitable or if you should try to make your existing business more efficient.
So the weak information learning paradigm would apply to H3 projects and the strong one applies to H1 projects. H2 would need the hybrid approach to avoid a Buridan’s ass dilemma.
The Hedgehog and the Fox
The Hedgehog and the Fox refers in an essay of the same name to the pattern where “a fox knows many things, but a hedgehog one important thing”.
Early Adopters and Laggards
Similarly, the dichotomy of early adopters and laggards fits well here, even if we just take the concept of learning (it does have a much broader context, though).
Early adopters would need to be always on their feet, ready to digest new and fresh ideas. Often it comes with both opportunities to seize on novel knowledge by extracting value from it before others. It also comes with risk as the raw ideas are untested and might incur a significant opportunity cost.
Laggards, even though they usually miss out on the highest yielding opportunities, what they do seize upon, usually hold value with high certainty. This is a much safer strategy and depending on the context, might be on average more valuable over time. For instance, in financial markets the boring proven investments (like large market index funds) have performed historically better than risky investments (like venture capital financing).
(I know I am conflating knowledge acquision and financial markets here, but I feel like it fits because investments follow knowledge.)
