Mental Models, Going Wide and Going Deep
Part 2
Take me to Part 1
I am happy to have you back for the second article in the series. I hope since you read the last article that you have been perhaps a bit more conscious of the mental models you hold in your own head, and how they represent what is going on around you.
In our last chat, we discussed thinking of real-world concepts in terms of the mental models we hold of them in our brain. We spoke of separating the fundamental, non-changing ones from the ones that are intrinsically possible to change. We also spoke of using these models to explain situations around us and to help break them down to fundamentals so that we can build them back up and change different aspects of them to create better outcomes.
The Wide Mental Model
In this article, I want to talk about what I call a broad or a wide mental model in contrast with a lower-level fundamental one. The wide mental model is one you hold in your mind about a larger domain such as computer science, photography, or perhaps baseball. It is your working knowledge of all the major subdomains, rules, and best practices that you associate with the subject. I have heard the analogy of a tree trunk used by Elon Musk. Your trunk is the base knowledge, and the leaves are the additional learnings you collect and fit on to the branch structure of your previously constructed wide mental model of the subject. Learnings without a wide mental model are quite a lot like leaves without a tree trunk and without branches in between.
You understand the new piece of knowledge in and of itself, and you may be able to use it in certain well-defined circumstances similar to the ones for which you learned it, but because there is no place to attach it to in the web of wide mental models that you already hold in your head, you can’t get much innovative use out of it.
Let’s take an example. Suppose you learn how to Spackle a hole in your wall. You open the Spackling paste, use a scraper to put it into the hole and smooth it out, and then you wait for it to dry and sand it. Simple.
You now have a leaf about how to fix a wall with Spackle. But likely you could never use it to innovate or draw much wisdom from it. You’d need to know the principals that make Spackle work, such as how it dries. You’d need to know the different kinds of Spackle and the techniques for different kinds of home projects. You’d also need to know the details of the wall itself. As you work toward a much lower level knowledge of Spackle, you’d get to the chemical behavior of the substances involved.
These details are where the models have the most usefulness. They are where the innovation can happen. You’ll rarely innovate by collecting leaves. You have to put in the time to develop strong trunks. A wide mental model is the beginning of a strong trunk. It is a 10,000-foot view of the domain itself and everything that falls within it. Over time, we fill them in as we learn and practice.
The larger number of wide mental models we have, the more breadth of the world around us exists in our heads to be used for deconstructing standards and drawing knowledge from different subjects. We do these deconstruction exercises to create new solutions.
Building Our Models
How do we go about building our wide mental model, and then how do we fill it in?
The first step is to scope the domain area by doing some initial research about it across a lot of different mediums to get different perspectives. Research it online, skim through a book at the library, maybe converse with a knowledgeable friend.
Then, divide it into sub-areas where each represents a different topic that is one of the most important in that domain. I think of it like a taxonomy or classification from biology that arranges and relates a number of different species.
First, you want to build your classification so that you can fit what you learn in the future against it. One place to find a good pre-built classification for your wide mental model is to look at a college curriculum. They nicely display the most important topics in a domain on one a handy document called a Syllabus. A quick look at a random culinary class shows the following subareas: Cooking vs. Baking, Kitchen Equipment, Cooking and Baking Recipe Vocabulary, Fresh Ingredients, Pantry Essentials, Wet and Dry Ingredient Measurements, Meal Prep, and Kitchen Safety. This provides a nice starting point for the novice to understand a few of the major areas of importance in the culinary discipline.
The classification will help you understand where to begin with deeper research. After we create the classification, how do we fill it in?
This is where you want to add hands on experience to your study. I use a mix of research, experiments, proof of concepts, and month-long projects to fill in mental models. Research without the hands-on experience makes a more brittle knowledge as we are all kinetic learners. We learn by doing in the real world. It’s easy to consume knowledge from different materials. It takes much more effort to practice and produce something that will challenge you to understand at a deep level. But the deep level is where the creativity happens, so we have no choice.
Study and Practice as approaches to filling a model creates a whole is greater than sum of parts effect on your knowledge of the subject. We want to create a good working knowledge of the high-level models with their classifications, and we want practice with projects to sink in the gritty details of the mental models that apply.
Multi-Disciplinary Insights
There were many major discoveries made throughout history that come from crossing knowledge between wide mental models. There are plenty of minor daily innovations that people create to make their day just a little bit easier as well. Let’s take a look at a few in the areas of business, accounting, and engineering.
Sharing Economy
One of the more obvious cross-discipline models in business, which has consumed the last decade, is the sharing economy. It has brought to life many services built around the coordination of sharing reusable items. This innovation required just two things.
One is that it required a way for a coordinated communication between many different parties. This was provided by the reliable internet and mobile phone technologies of the 21st century.
The other is the recognition and then social acceptance that we can share our personal assets with strangers while we are not using them ourselves.
With these two concepts in place, most of anything reusable can turn into a sharing system. We have Uber for driving, CityBike for biking, Airbnb for house-staying. It is easy to see how these designs evolve to become rule in our society, as opposed to just one out of many options in the potential space of innovations that was tried successfully and stuck. The sharing model can easily extend to fishing poles, snow shovels, and jet skis, if we like.
Double Book-Keeping
Another model that can be crossed into lateral disciplines is the idea of double-bookkeeping in accounting. It is an accounting system that involves balancing account totals in such a way that money or supplies leaving the business is equal to the money or supplies coming in. You keep track of totals in two places and they always have to be equal. If they are not, then a record of a transaction is incorrect.
The important part is that you are checking twice against yourself for any human error that may have been made. The idea of checking something twice in two different ways can reduce error greatly.
Suppose you are counting your red, blue, and green socks on laundry day to make sure that you have them all. You can check that you have all socks in two ways by counting all of them together and then by counting each of the different colors individually, adding together the totals, and comparing. There is a reasonable chance you may make a mistake the first-time accounting, especially if you have a lot of socks. But it’s very unlikely that on the second count, you would make the exact same counting mistake and again arrive at the same incorrect number. You might get a new incorrect number, or you might actually get the correct number, both of which would tell you to recount.
Double counting for reconciliation purpose is a powerful accounting tool that applies to many different disciplines.
Margin of Safety
Let’s take one more example. There is a concept in Engineering called a Margin of Safety. It is used in bridge design and likewise in investment analysis. It is another model favored by Charlie Munger. It’s about over-preparing to give yourself a reasonable amount of leeway in case some unfortunate issue occurs. It can be as simple as giving yourself just a little more distance than needed from a swerving car, or wearing something slightly warmer than you think you’d need for the day’s weather.
Good margins of safety are balanced. They create enough distance to survive the most reasonable mishaps, but not enough to be overbearing in the case of events that are highly unlikely to occur. Wearing multiple layers on a warmer day for instance is probably more distance than is reasonably needed and crosses into overbearing territory.
I’d recommend getting into the habit of thinking of margins of safety in daily actions. They have saved me both in smaller and larger circumstances more than a few times. Like double bookkeeping, a margin of safety can be applied to many different subjects.
How does someone recognize the cross-application possibilities found in the examples above?
Many cross-domain solutions like the examples above come in the form of AHA moments when we are doing something else besides focusing or studying. Ordinarily, people are not thinking in a cross-domain fashion because they would not study two very disparate subjects like cooking and economics in the same sitting. More likely, they come in a less focused state when we are doing something else, and we are thinking about many different subjects at random.
We will talk a little more about this state in the next article of the series which will describe how the brain learns, and how it creates these innovative AHA moments.
For now, know that cross domain innovation comes from having multiple wide mental models developed and accessible.
Model Reconsolidation
As we fill in mental models, from time to time there will need to be a reconsolidation of the material. This will involve stepping back to see the big picture and incorporate all recently learned material into your overall knowledge of the subject.
Have you ever picked up a new piece of information but just couldn’t seem to fit it in with what you already know? You then have to spend time thinking and consolidating the new with the old. You have to adapt your mental model.
When I started this, I did these reconciliations with no apparent schedule or methodology. Most of the time, it would happen in between focused study like at a doctor’s office. I have since found a more structured way to fill my wide models with new information.
Let’s draw from another domain and cross it with what we are learning here.
Version Control
There is a very well-defined industry standard in the area of software development called versioning. When somebody builds the code for a software product, like a website, that person goes through a process of creating the code, packaging it, and pushing it to the server that “delivers” the site to people’s computers. Every time an upgrade is needed on the website and a new version of the code is produced, it follows the exact same process of packaging and being pushed to the server. Now, perhaps after pushing it to the server, the developer realizes there is a mistake in the code and the website is no longer usable. He now wants to quickly revert back to the previous version of the code and try again to release the new version later, once he fixes it.
To do this he needs to know which previous version he wants to push to the server. He needs some kinds of annotation system; he needs to annotate the versions with a serial number of sorts and maybe a description so he can reference them later. Many standard software tools which help with packing and pushing new versions of code include a build number by default to help with this versioning requirement.
What I’d like you to take from version control is the concept of versioning work and the insight that it can be applied to learning and practice as well. I apply it to my mental models.
After a period of learning new information, I schedule some time to work it into my previous understanding of the subject as a whole. I periodically step back to rewrite and redraw my mental model in new draft formats with a version number. This lets me see over time how my working mental model of the subject changes and ultimately gets better and closer to the real-life concept that the model represents. The trail of drafts also shows what became most important about the subject over time and what became less important.
How We Innovate
We build our wide mental models by filling them in with lower level models and best practices as we research, practice and experiment. Having a deep knowledge of a few different broad mental models will best help us draw knowledge together from each of them to build better solutions.
Furthermore, all wide mental models can be broken down into the fundamental mental models, a concept we had introduced in the previous article. There are many unchanging models in the areas of psychology, economics, and engineering upon which the higher-level wide mental models are built. They are unchanging because they are built into either unchangeable or very slowly changing aspects about our world, such as the economics of materials and matter or the psychology of the human being. Both change very slowly compared to the fast pace we innovate the human-built systems around us.
Breaking down a higher-level model into the fundamentals while maintaining an understanding of the purpose of that model can help us understand what parts of it cannot change and what parts of it can. We use the changing parts to understand what to improve upon to better serve the purpose. This is how we “do” innovation.
In the fourth article of the series, we will break down some of these high-level models. The next article will digress from our main focus of the mental model strategy to understand how the brain best learns and thinks.
Over the next few days, make some time to take a look at the subject or domain you are most familiar with. Try to break it down into sub-domains and the fundamental subjects that it may rely on to work.
Part 3 of the series will discuss building mental models and the nature of cross-discipline insight. If you missed Part 1 on the innovation strategy, be sure to have a look.
Thank You for reading.
Glossary of Terms and Relationships
A Subject or Domain is a broad area of study that groups like knowledge together like Economics, Sports, or FishingA Wide Mental Model is a classification of important sub-areas within a subject or domain, it contains subdomains, rules, and best practicesA Sub Area is a lower-level mental model within a subject that holds some major important for those knowledgeable about the subject. The importance could be based on usefulness, interest, or necessity. The area of tax study under the subject of accounting is an example of an important one because it is useful to many. The virtual reality sub-area in the video game domain might one of great interest to somebody, and so it would earn a place in their wide mental model for video games. Air resistance mathematics is a necessary part of airplane and flight design, and so it might earn an important spot in the mental models of the airplane designer or enthusiast.Model Reconsolidation is incorporating all recently learned material into your overall knowledge of the subject you already hold, it is consolidating your new knowledge with your old knowledge. It can be versioned to structure your progress.
Citations
[1]: Charlie Munger,(October 3, 2003). Academic Economics: Strengths and Faults After Considering Interdisciplinary Needs https://www.youtube.com/watch?v=jY1eNlL6NKs
[2] Elon Musk, (Jan 5th, 2015). https://www.goodreads.com/quotes/5444442-it-is-important-to-view-knowledge-as-sort-of-a
