Some time ago, someone introduced me to the terms “lumpers” and “sorters”, and that stuck with me. They describe extreme ends of a spectrum.
- Lumpers:
- Lumpers are people who prefer to see commonalities between things, and so “lump” them into the same category.
- Sorters:
- Sorters by contrast prefer to focus on differences, and so sort things into highly specialized categories.
Reasons for Lumping and Sorting
I find this pretty fascinating. I’ve been accused of lumping far more in life than I’ve been accused of sorting, which at a cursory glance might mean I have a clear place on that spectrum. I don’t disagree, but I also don’t fully agree.
What I observe, of course, is that most other people seem to be sorters. Obviously, on might think, when I’m a lumper.
There’s a reason I veer towards lumping, though, and I don’t seem to have the same reason other people appear to lump. This relates somewhat to the reason I see other people seemingly prefer sorting.
Now when I speak about other people’s reasons, that’s a pretty haphazard thing to do. In almost every case, I will be wrong. But I’m lumping here, in my own way, not sorting, so bear with me a little. Try to focus less on the ways in which I am wrong (I know I am), but on the ways in which I am right (I know I am that, too).
Very broadly speaking then, I see people have different priorities when seeking solutions.
- It appears to me as if lumpers have a preference for fast solution finding. In the pursuit of speed, they appear more willing to overlook “small” differences, because it is easier, and so faster, to find a solution matching the problem at hand.
- Conversely, I get the impression that sorters have a preference for the correct solution. It seems they sort in order to be certain that the solution they choose matches the problem as best as possible.
I don’t know if this image helps anyone but myself, but imagine that solutions to a problem “branch out” into a tree-like structure in front of you. At every fork, a different detail of the problem presents you with one or more new choices of paths.
In this image, sorters try to traverse all the branches of the tree before settling on one. By contrast, lumpers “collapse nodes”, i.e. prune branches that seem to lead to unnecessary distinctions, until a clear path is left.
Cognitive Dissonance
I’m struck time and again by how much cognitive dissonance appears to play a role in human decision making. Put briefly, the theory states that this dissonance in one’s cognition is unpleasant, so we seek to reduce it. The dissonance arises when we observe or experience something that clashes with something we already believe or know.
There are various ways of doing so. One of the simpler and more harmful things people do is reject new information as false when it disagrees with already learned things. But the opposite way, to question the learned things and incorporate the new information would also reduce dissonance – but it would take longer and might be more difficult.
If you want to couch problem solving in terms of cognitive dissonance, I would say that the dissonance is caused by different things.
For lumpers who prefer fast solutions, it would appear that the existence of the problem in the first place has an outsized weight on the dissonance. Getting the problem out of the way is more important than the specifics. “To do nothing is the worst.”
Meanwhile, sorters might take the problem more as a stimulus. The real dissonance arises from imagining the damage of applying the wrong solution to a problem. “To do the wrong thing may be worse than to do nothing at all.”
I don’t know if this characterizes lumpers and sorters well.
When Lumping is not Lumping
I don’t really like being put into the “lumper” category, because another thing you’ll see me do if you observe me for long enough is to prefer a slower more roundabout route to a solution if it solves the problem better.
Why do I write this? Well, my experience in conversations with sorters about why I shouldn’t lump is that sorters, in applying highly specific solutions to problems, have a tendency to treating problems and their corresponding solutions as unrelated.
Your shirt button popped off? Well, sow it on. Your belt feels tight? Loosen it. Maybe that’s all there is to be done. But maybe you’re hulking out, and that’s the problem you should really focus on.
What I do when I “lump”, is I do try to prune unnecessary detail, for the moment. What is or is not unnecessary depends. For everything I prune, I try to find a solution. The solution, though, needs to come in two parts:
- First of all, the solution must solve the common path up until the point where I pruned.
- Second, if I add the pruned branches back, then a minor addition to the solution should address each of the detail branches well.
I iterate over this with different things pruned out.
Doing this is hard work, and sometimes doesn’t provide a perfect solution. There are times, when specialized solutions for specialized problems are the only path to take. Or, more specifically, if this approach does not provide a good solution, then there is no choice left but to sort.
But when it works, well. That’s when solutions become elegant. You’ve pruned enough to show the true, underlying shape of the problem, and solved that.
Complexity and Complication
I think my own source of dissonance is complication.
When having the terms “complication” and “complexity” compared, you can find simple definitions: complex means a lot of stuff is happening. Complicated means stuff is hard to understand.
A more nuanced but harder definition contrasts “complicated” to “simple”. Both describe classes of predictable problems, which therefore allow for a controlled reaction to the problem, a solution design.
By contrast, complex problems are harder to predict, which means solutions may not be readily apparent. Perhaps solutions are less about a designed reaction that controls the situation.
Here’s the problem with “complicated”, from my point of view: you start out with a simple problem, and apply a simple solution. Another problem arises, and you apply a simple solution. This continues, until you have a complicated collection of solutions which address a complicated problem landscape.
You find a new problem. You apply a simple solution. Some seemingly unrelated part of the solution realm collapses. Unpredictably.
Your complicated solution landscape suddenly has become complex.
I’m not actually afraid of complexity. But often enough, what appears to be complex is just a complicated thing that has outgrown your ability to predict. (It’s still theoretically predictable; you just can’t handle it any longer.)
I don’t like lumping, because it will never fit everything, in the worst case to the point where it won’t fit anything. I don’t like sorting, because it so often leads to complexity emerging. The world trends towards entropy.
“Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.
― Antoine de Saint-Exupéry, Airman’s Odyssey
I do my lump-then-sort approach, because when it arrives at solutions, it tends to arrive at relatively simple solutions for a relatively complex problem landscape. By allowing for a basic, shared solution shape with variable extensions for addressing detailed problems specifically, it has a better chance at withstanding the eventual collapse into entropy, or at least lasting longer.
Techniques
Here, I strafe programming a little. There are a few things that one can do to helping keep a solution landscape maintainable, and flexible for future extensions.
Decompose the problem. Most problems are themselves a system of smaller problems interacting. Those smaller problems are often simple, and often have simple solutions. It’s perfectly fine to apply those simple solutions to those simple problems; it’s the composition of those solutions that solves the apparent problem.
Given that this composition is so important, that is what the lump-then-sort approach needs to focus on. This needs to be kept simple and flexible.
“Duck typing” is fine. “If it walks like a duck and quacks like a duck, then in this specific solution context it’s OK to treat it like a duck.”
This isn’t lumping as such. It’s saying that if it’s a door or a gate doesn’t matter when the problem is allowing entry at a given time, so the solution needs to involve having someone with the key around.
When the problem is what paint to paint it with, the difference may be immense.
Related to this, allow for different filter lenses on the same problem. A green filter and a red filter will discard different detail. Sometimes what remains reveals the underlying shape of the problem. Sometimes it reveals something you didn’t consider.
I think the upshot here is to give “lumping” a try, as long as you don’t just lump once in pursuit of a speedy solution. Lump with abandon, and in different ways. You’ll learn a lot about the problem.
Then try to make the basics simple. Time will reward you.
I’m not sure if this is a helpful thing for you to read. It’s one of those topics that I write down for my own benefit.
“For my own part, I have never had a thought which I could not set down in words, with even more distinctness than that with which I conceived it.”
― Edgar Allen Poe
Writing things down sometimes helps me with achieving greater distinctness.
In this case, it’s that someone else finally gave me the words “lumpers” and “sorters” to describe the spectrum, and that allows me to write about how I feel I’m neither, or both. The topic has followed me enough through life that writings this down is more of an attempt to settle it. I never knew how much I needed those terms.
My cognition experiences dissonance when my thoughts are not sufficiently distinct to be able write them down.