Reasoning from First Principles
Make the implicit explicit
With the abundance of information available these days, particularly that which is found in punchy articles, I’ve found myself frequently in a familiar situation. I’ll be reading along, convinced by the claims made by the author until I get to the end. The conclusion — which often promises health, financial independence, improved productivity, or career success— is dubious. This happens when my definitions or assumptions are different from those of the author. For example, the wealth required to be financially independent in a low cost of living area is much lower than that required for a high cost of living one, or the career advice offered may be applicable to folks at large companies but not small ones.
Sharing explicit definitions and assumptions up front benefits the reader, as it gives them an idea if the conclusion will be well suited to their situation. This is known as reasoning from first principles, and has been used by great thinkers such as Johannes Gutenberg and Elon Musk. A first principle is a definition or assumption that cannot be derived from other definitions or assumptions. I was first introduced to this style of thinking in math class, where definitions and assumptions (in math, postulates) are relied upon to derive conclusions (in math, theorems). This results in a logically consistent system, where no one can dispute the resulting theorems as long as they agree with the original definitions and postulates that the system is built on. Although the real world lacks the certainty of mathematics, we can borrow these techniques to strengthen our arguments and advice.
I realize that too much space dedicated to definitions and assumptions can turn off readers before you can share the crux of your work with them. However, I do believe there are artful ways around this problem:
- One can rely on definitions and assumptions that others, particularly highly regarded folks in your area of interest, use. This forces you as a writer to align your definitions of terms and concepts with those of the best thinkers in a given space, and if your reader isn’t familiar with these thinkers or definitions, you’ve taught them something valuable at the beginning of your work! This feeling encourages them to continue reading your piece.
- Alternatively, one can define terms and outline assumptions oneself. This is a difficult exercise, and one that most folks shy away from for that reason. However, doing this effectively can impress readers, particularly when you can come up with definitions or assumptions that differ from conventional ones in inspired and defensible ways. Additionally, defining terms or sharing assumptions doesn’t always have to add length to your pieces, as you can house them in separate works that you reference in subsequent ones. By doing so, you convert what would be individual, vague pieces into a strong body of work with a set of obvious, common themes. This helps you lean into Medium’s recent announcement about encouraging relational media consumption. Your comprehensive understanding of topics from definitions and assumptions to conclusions encourages readers to repeatedly engage with your work.
Although I typically read and write articles in the data science space, and am guilty of not defining terms and concepts before my first blog post in that space (to be remedied soon!), I’m going to use a less technical and hopefully more accessible example to illustrate the importance of making definitions and assumptions explicit as well as the surprising finding that first principles reasoning encourages creativity.
Definitions
Imagine that you have some spare time and decide to pick up cooking as a hobby. You’ve always had an interest in it, and are able to throw together a few simple dishes, but you wouldn’t consider it a skill you’re particularly good at yet. So, you decide your goal is to become a good cook.
What’s your first step in this journey? Some folks, excited with having chosen a new hobby, would get right to it and whip something up with whatever they have on hand. Others would read highly reviewed cookbooks. Some would start buying fancy kitchen gadgets that they’ve seen chefs on TV use. Although all of these strategies seem reasonable, we’re doing them before we’ve even defined what being a good cook actually entails! What if being a good cook just means preparing a dish we’re proud of? Does that require spending hundreds of dollars on new kitchen gadgets that we just bought?
Before we do anything, we should define what being a good cook means to us. Are we a good cook if we win a cooking show like Top Chef or Chopped? Do we need a James Beard award or a Michelin star? Perhaps we decide that we don’t need public accolades, we’d be able to accomplish our goal if our close group of friends viewed us as a good cook. Although it may feel like an unnecessary exercise, we could also define what qualifies as cooking. Does cooking include preparing salads, baking bread, or making desserts? One could argue for, against, or even take a middle ground (e.g. preparing desserts isn’t necessary to be a good cook, but it helps). One can see the parallels between our cooking example and articles on becoming a “great data scientist” or “top investor”. Not only should the author define the subjective “great” and “top”, but also the more objective fields of “data science” and “investing”. Does the former include data engineering and communicating findings to business leaders? Does the latter include short term trading or holding more speculative assets?
Assumptions
Since we’ve identified our goal: being thought of as a good cook by our friends, we need to figure out how we’re going to achieve it. With the goal in mind, it’s obvious that we should cater to our friends’ palates. We should ask ourselves which cuisines and specific foods do they prefer. Do they want a more rustic or refined style? Do aspects of cooking like plating and the use of cutting edge techniques matter?
Better yet, we could strengthen our assumptions by asking them these questions directly. Why rely on our memory of dining experiences with them when we could go straight to the primary sources? Some would argue that by doing so, we turn what were assumptions (our impressions of what they prefer) to facts (what they actually prefer). I’d argue that what we’re doing is strengthening our assumptions. Why? Although we reduce uncertainty by asking them these questions, we don’t eliminate it. What if the questions aren’t framed properly, and they tell us that they prefer rustic food — but are actually more impressed by haute cuisine. Even if we ask them explicitly, “what would I need to cook for you in order for you to think I’m a good cook?” — which is probably awkwardly forward of us — their answer to that question could change between when we we get their answer and when we cook for them.
Encouraging Creativity
It was counterintuitive to me at first that a first principles reasoning framework could encourage creativity. I would expect a logical framework with explicit definitions and assumptions to help adherents come up with well justified— but unimaginative — thinking, but I was surprised to find that it could also lead to creativity. I believe this happens because without first principles reasoning, people overly rely on heuristics, which are rules of thumb that have broad applicability but are not appropriate for every situation. Heuristics are a common way of thinking because they take little effort and work well in most situations, but may not be optimal, or even work particularly well, depending on the specific situation you’re in.
In our example, before we defined our goal of being perceived as a good cook by our friends, we considered whipping up ingredients in our fridge, reading highly reviewed cookbooks, or buying expensive kitchen gadgets that we’ve seen chefs on TV use. Although all of these seem logical, they rely on heuristics and may not be appropriate or optimal for our situation. Below, I’ve outlined the proposed action, the underlying heuristic and possible flaws in our reasoning:
Action 1: Cook whatever ingredients we have on hand.
- Heuristic: In order to become a good cook, we need to practice cooking.
- Possible Flaws: If we have just started getting into cooking, it’s very likely that we have limited ingredients on hand. We might only be able to make very simple recipes which are unlikely to impress our friends.
Action 2: Read highly reviewed cookbooks.
- Heuristic: Good cooks should be able to make recipes from highly reviewed cookbooks.
- Possible Flaws: Just because a cookbook is highly reviewed does not necessarily mean that our friends will enjoy its recipes. Their tastes may differ from those of the cookbook author.
Action 3: Buy expensive kitchen gadgets we’ve seen chefs on TV use.
- Heuristic: Good cooks should use fancy kitchen gadgets to make their food.
- Possible Flaws: Our friends may not be familiar with or interested in food prepared using fancy kitchen gadgets.
So, where does creativity come into play? Recall that our goal is to have our friends think of us as a good cook. From having dined at restaurants with them, we already know of various dishes they like. A creative solution to showing off our cooking skills (after we’ve practiced them, of course) would be to serve some of these well liked restaurant dishes next to ours when we have our friends over. If our cooking is as good as, or even better than, food we know they like, that could make our newly developed cooking skills more apparent to them. You may note that ordering food we know our friends like to compare our dishes to doesn’t directly involve cooking, unlike all of our heuristics based actions. One can devise unconventional solutions like this by staying focused on definitions and assumptions instead of overly relying on heuristics.
Because of the extra effort required to do first principles reasoning, I’m certainly not advocating for folks to do it in all areas of their lives. And even for the cooking example I shared above it feels like overkill; hopefully you enjoy the act of cooking in addition to being able to impress your friends with what you make. However, I do think it is important and necessary when writing or thinking about topics that are most important to you, and in particular when you’re giving advice to others. It strengthens your understanding of concepts in your area of interest, keeps you focused on your goals, and encourages creative solutions.
