There's a baby carrier sitting in our hallway that hasn't been touched in six months. It's not broken. It's not uncomfortable. It's just... incomprehensible. Daniel wrote in about this, and I think half the parents listening just nodded involuntarily.
He's strong, he's handy, he can rebuild a computer and drill a straight hole through a wall stud. But hand him a harness with six adjustment points or ask him to tie a bowline, and his brain just stops. He wants to know what cognitive faculty that is — what specific slice of intelligence handles knots, levers, straps, and mechanical systems — and why it can be missing while other spatial skills are perfectly intact. He's also wondering whether his ADHD plays into this, because the pattern he describes is familiar: the instructor says move this part over that part, and he has to slow everything down and say show me which part and show me what other part.
So today we're going to look at what the research actually says about mechanical reasoning, spatial visualization, and why your brain can be great at some spatial tasks and utterly flummoxed by others.
The thing Daniel is describing has a name in the cognitive literature, and it's not just spatial ability. It's mechanical reasoning — sometimes called mechanical comprehension. And it's a surprisingly narrow thing.
Let me pin down what we're actually talking about here. Mechanical reasoning is the ability to understand how physical systems work — how forces transmit through levers, pulleys, gears. How flexible objects like ropes and straps behave under tension. How the order of operations changes the state of the system. It's not the same as being able to imagine a shape rotating in your head, and it's not the same as knowing which end of a screwdriver to hold.
So Daniel drilling a hole — that's one thing. Daniel trying to figure out which strap to tighten first on a baby carrier so the whole thing doesn't slide sideways — that's a different thing entirely.
No, wait. Let me rephrase. The CHC model of intelligence — that's the Cattell-Horn-Carroll model, which is the most widely accepted framework for organizing cognitive abilities — it separates visual processing into a bunch of narrow abilities. Spatial visualization is one of them. That's Gv-Vz. It's what you use to mentally rotate a shape or imagine what a 3D object looks like from a 2D blueprint. Daniel's DIY skills and computer repair suggest his spatial visualization is strong.
But mechanical reasoning — sometimes labeled Gv-MK, sometimes placed under fluid reasoning as Gf-MK depending on which version of the model you're looking at — is a separate narrow ability. It involves understanding dynamic physical interactions, not just static spatial relationships.
So the CHC model already anticipated that someone could be good at one and bad at the other. It carved them into different buckets.
It did. And here's where it gets really interesting for Daniel's question. There was a meta-analysis in 2023 — Sorby and colleagues — that looked at whether training people in spatial visualization transfers to mechanical reasoning tasks. It doesn't. You can get better at mental rotation all day long, and when you sit down in front of a pulley system or a belt drive, you're still starting from scratch. That suggests these are distinct neural substrates. Different brain circuits doing different work.
Which means Daniel's experience isn't a contradiction. It's not weird that he can route cables inside a computer case — which is a spatial puzzle with rigid components — and then freeze up on an Ergobaby carrier. The carrier is a completely different kind of system.
This is the distinction that I think explains most of the frustration. A computer case is a rigid system. The components don't change shape when you move them. A strap or a rope is a compliant system — it changes shape under load, it has multiple degrees of freedom, and the state of the system depends on the order of operations. Which strap you tighten first changes how all the other straps behave.
A hole is a hole. You drill it, it stays drilled. The drill doesn't care what you did before or what you'll do after.
Right. And a knot — take a bowline, which is one of the most useful knots there is — a bowline requires the rope to be formed into a specific topology. A loop with a particular crossing pattern. You cannot understand a bowline by looking at the finished knot. You have to understand the sequence of transformations that produced it. The rabbit comes out of the hole, goes around the tree, and back down the hole. That's a procedural mental model, not a static image.
The rabbit and the tree thing — I've heard that mnemonic. Does it actually help?
For some people. For others it's just more words to hold in working memory while they're also trying to hold the rope. And that's where the ADHD connection comes in.
Let's go there. Daniel mentioned it specifically.
ADHD is associated with working memory deficits, and working memory is exactly what you need when you're building one of these procedural mental models. Understanding a knot or a harness requires you to hold multiple spatial relationships in your mind simultaneously while executing steps in sequence. You have to simulate how forces propagate through a flexible network. That's cognitively demanding, and it relies on mechanical comprehension — a skill that, by the way, correlates with general intelligence at about r equals 0.4 to 0.5. Moderate, not strong. There's a lot of individual variability.
There was a study — Martinussen and colleagues, 2021 — that looked at children with ADHD and found specific deficits in mechanical reasoning tasks that required holding a sequence of spatial transformations in mind. Not in all spatial tasks. Specifically the ones where you have to do step one, remember what step one did to the system, then do step two while keeping the result of step one active.
That's the instructor saying move this part over that part. Daniel's brain is saying I need to see which part and which other part, because the words alone aren't enough to build the model while also executing the action.
And if you're in a sailing class with eight other kids and the instructor has already moved on, you're not just confused — you're also feeling like the slow one. The emotional layer on top of the cognitive difficulty is real.
So let me see if I've got the mechanism right. The compliant system — ropes, straps, harnesses — demands that you build and update a dynamic mental simulation. That simulation lives in working memory. ADHD constrains working memory. So the system that's hardest for Daniel is the one that hits his cognitive bottleneck hardest. The rigid-system tasks he's good at don't tax working memory in the same way.
That's it. Drilling a hole is linear. You mark the spot, you drill. The workpiece doesn't change its properties mid-operation. A baby carrier has — what, six adjustment points on some models? You tighten the shoulder straps, and suddenly the waist belt is sitting differently. You fix the waist belt, and the chest clip has migrated up to the kid's throat. The system is coupled. Everything affects everything else.
I've watched people put on those carriers. There's a moment where they're just standing there with straps dangling everywhere, rotating the thing around like they're trying to solve a Rubik's cube by feel.
And some of them give up. Daniel mentioned he's avoided using the carrier because he can't remember the sequence. That's not laziness. That's cognitive load avoidance. It's a completely rational response to a task that costs more mental energy than it returns.
The carrier sits in the hallway. The kid gets heavy. The parent's back hurts. All because the instruction manual assumes a cognitive faculty that not everyone has in equal measure.
This is the knock-on effect that doesn't get talked about enough. It's not just a party trick — can you tie a knot or can't you. It's car seat installations that get done wrong because the belt path through the seat frame is a compliant system problem. It's flat-pack furniture with tension straps that never gets assembled. It's people who reach adulthood knowing exactly one way to tie their shoes and nothing else.
I know one knot. It's... not a good knot.
Most people's one knot is just several overhand knots stacked on top of each other and hoping for the best.
That's the one.
So what actually helps? Because the research on this is useful, and it points in a direction that's counterintuitive.
Go on.
Most instructions for knots and harnesses are step-by-step. Do this, then this, then this. Photos with arrows. Videos where someone's hands move through the sequence. That approach works for people whose working memory can hold the sequence. For everyone else, it's a recipe for getting lost at step three.
There was a study in 2022 — Leopold and colleagues — that tried something different. Instead of teaching knots as a sequence of steps, they taught the topology. The pattern of connections and crossings. They used color-coded ropes and freeze-frame diagrams that showed what the knot should look like at each stage — not what your hands should be doing, but what the rope should look like. Performance improved by sixty percent compared to step-by-step video instructions.
Sixty percent. That's not marginal.
It's enormous. And the insight is that the topology approach offloads the working memory demand. You're not holding a sequence in your head — you're comparing what's in front of you to a reference image. Does my rope look like the picture? No? Then I'm not at the right stage yet.
So it's the difference between memorizing a dance routine and having a mirror in front of you.
For harnesses and carriers specifically, there are a few strategies that the research supports. One is to decompose the system into rigid sub-assemblies. Find the parts that don't move relative to each other and treat them as a single object. On a baby carrier, the main body panel and the waist belt might be one sub-assembly. The shoulder straps are another. Figure out how they connect before you try to adjust anything.
Treat it like a wiring diagram before you touch any wires.
Another technique — and this is so simple it sounds stupid, but it works — is the one hand always holds the reference point method. Keep one finger on the anchor point. On a carrier, that might be the point where the shoulder strap meets the back panel. As long as you're touching that, you can't get completely lost.
That's a proprioceptive anchor. You're using your body's sense of position to supplement the mental model.
And the third one, which is the one I'd recommend to Daniel specifically: lay the carrier flat on a table and study it before you put it on. Understand which strap connects to which anchor point. Not which order to pull them — just the topology. This strap goes from here to here. That one crosses under this one. Build the static map first, then add the dynamics.
The sequence is the last thing you learn, not the first.
Which is the opposite of how every instruction manual on earth is written.
Instruction manuals are written by people who already understand the system. They're the worst teachers.
They really are. And this connects to something broader about how we think about spatial intelligence. Most spatial ability tests use rigid objects — blocks, shapes, puzzles. The classic mental rotation test is a picture of a 3D shape made of cubes, and you have to figure out which of four options shows the same shape from a different angle. That's a rigid system. It doesn't change. The real world is full of compliant systems, and we barely test for the ability to reason about them.
So people grow up thinking they're bad at spatial tasks because they can't tie knots, while simultaneously being good at every spatial task that involves rigid objects. The tests don't capture the thing they struggle with, so they never get the feedback that says this is a narrow gap, not a general deficit.
And they feel stupid. Daniel's sailing instructor probably thought he was challenged, in his words. But the instructor was teaching a compliant system as if it were a rigid one — follow these steps, in this order — and Daniel's brain doesn't work that way.
The ADHD angle here is interesting because it suggests the fix isn't to try harder at the thing that's hard. It's to route around the bottleneck.
Executive function supports. Externalize the sequence. Use checklists. Color-code the straps — you can buy little colored tabs that clip onto webbing. Take a photo of the carrier when it's correctly adjusted and keep it on your phone. The key insight is that the difficulty is not in understanding the physics — it's in holding the sequence of transformations in mind while executing them. Offload that to the environment.
So Daniel's not bad at straps. His working memory is bad at holding the strap-sequence simulation. Different problem, different solution.
And here's the thing — this cognitive variation is completely normal. Mechanical comprehension has high individual variability. Some people are at the ninety-fifth percentile, some are at the thirtieth, and most of them have perfectly intact general intelligence. Daniel's DIY success proves his spatial visualization is fine. The gap is in this narrow ability that we don't even have good names for in everyday language.
We call it being handy, which conflates about six different cognitive skills into one word.
And then we're surprised when someone who's handy in one domain isn't handy in another. It's like saying someone who's good at sprinting should also be good at swimming because both involve moving through space.
The comparison to learning a musical instrument keeps coming to mind. Some people play by ear — holistic, intuitive, they just feel it. Others need sheet music — sequential, explicit, they need the map. Both paths produce musicians. But if you hand sheet music to the ear player, they look incompetent. And if you take away the sheet music from the sequential learner, same thing.
I think that's exactly what's happening with Daniel and the sailing knots. The instructor was teaching the sequential path — move this part over that part — and Daniel needed the topological map. Or maybe he needed the kinesthetic path.
Kinesthetic?
Learning by feel. Proprioception and tension feedback instead of visual-spatial reasoning. Some people tie knots by understanding the physics of how the rope wants to move, not by memorizing a diagram.
That's... actually where I want to bring in someone who has a very different relationship with this topic.
Hilbert: I think you're overcomplicating it.
Hilbert: For me, knots are about tension and friction. Not topology. I never think about which part goes over which part. I just pull it tight and if it feels wrong, I start again.
Hilbert: I worked three summers as a climbing wall instructor at an outdoor center in the Lake District. Keswick. I taught people how to tie figure-eight follow-throughs, how to attach harnesses, how to set up belay systems. I was the weird one because I could tie knots behind my back but I couldn't explain how. I did it by feel.
Hilbert: The problem with Daniel's approach — and with what you two have been describing — is he's trying to understand it like a diagram. Straps and ropes don't care about diagrams. They care about load paths. You can tie a bowline wrong and it'll still hold if the load is right. You can tie it perfectly and it'll slip if the rope is slick.
Hilbert: My brother-in-law ties knots for a living. Commercial fishing out of Ullapool. He ties maybe two hundred knots a day, has done for thirty years, and if you asked him to draw a bowline on paper he couldn't do it. His hands know it. His eyes don't.
Hilbert: He also says — and this is the part you should probably not take as gospel, because the man has been thrown off a pier twice — he says the people who struggle most with knots are the ones who try to understand them before they touch the rope. They want the mental model first. Then they pick up the rope and it doesn't match the model and they freeze. He says you have to let the rope teach you. Which sounds like something a man who's been thrown off a pier would say.
So there's a kinesthetic pathway that doesn't run through spatial visualization at all.
Hilbert: That's what I'm telling you. I can't visualize a knot. If you ask me to picture a figure-eight in my head, I get nothing. But my hands know the shape. They know when the tension is right. I could tie one in the dark. I have tied them in the dark. On a wall in the rain with a client panicking thirty feet below me.
Hilbert: The instruction manuals never tell you this part. They show you arrows and numbered steps. Nobody says close your eyes and feel where the rope wants to go.
That's a completely different cognitive strategy. You're using proprioception and haptic feedback instead of visual working memory.
Hilbert: I don't know what those words mean. I just know I can tie a knot and I can't draw one.
Hilbert: The other thing my brother-in-law says — and again, pier, thrown off, twice — is that the people who get good at knots stop thinking about them as shapes and start thinking about them as forces. This part pulls against that part. The friction holds here. The load runs along this line. It's not a picture, it's a... he calls it a tension map. He says you can feel a bad knot before you see it.
For Daniel, the advice might not be find a better diagram. It might be stop using diagrams entirely and switch to a tactile strategy.
That's interesting, because the Leopold study I mentioned found that topological diagrams helped, but Hilbert's describing a pathway that bypasses visual representation altogether. It's possible there are multiple alternative routes to competence, and the key is matching the route to the person's cognitive profile.
Hilbert: I'll tell you what doesn't work. Standing there with the instruction booklet open while the baby's crying and your partner's saying haven't you figured it out yet. That's not a cognitive problem, that's just... life. But it makes everything worse.
Hilbert: I still have a figure-eight tied in a piece of climbing rope from my last day at the center. It's in a box somewhere.
If you take one thing from this conversation, it's that struggling with straps and knots is not a character flaw and it's not a sign of low intelligence. It's a mismatch between a specific cognitive demand — building and updating a dynamic mental model of a compliant system — and the working memory resources you have available. Daniel's brain is great at rigid spatial tasks. It chokes on compliant ones because they hit a different bottleneck.
The fix isn't to get better at straps. It's to find a strategy that works for your brain. That might be topology diagrams instead of step-by-step instructions. It might be color-coding and checklists. It might be closing your eyes and letting your hands learn the tension map. The goal isn't to become someone who's good at knots. The goal is to get the baby carrier on and get on with your day.
There's an open question here that I keep coming back to. If mechanical reasoning for compliant systems is a narrow ability that doesn't respond to general spatial training, what does that mean for how we design instructions and products? Should baby carriers come with topology diagrams instead of step photos? Should car seat manuals include a section that says here's what the belt path looks like from three different angles, with the anchor points highlighted, before they show you a single step?
The evidence says yes. And it also says we should stop pretending that following instructions is a single skill that everyone has equally. It's not. It's a collection of narrow abilities, and some of the most common instruction formats systematically disadvantage people whose cognitive profile doesn't match the assumptions of the person who wrote them.
To Daniel and everyone who's ever stood in front of a half-assembled piece of furniture with a strap they can't figure out, or given up on a baby carrier because the harness defeated them — you're not alone, and you're not broken. Your brain just works differently. That's not a consolation prize. It's a fact about cognitive architecture, and it means the problem is solvable once you stop trying to solve it the wrong way.
Thanks to Hilbert Flumingtop for producing, and for the reminder that some things are better learned in the dark with cold hands and a rope that doesn't care about your diagram.
This has been My Weird Prompts. If this episode resonated — and I suspect it did for a lot of you — we'd love to hear about it. Email the show at show at my weird prompts dot com.
We'll be back soon.
