In my first post in this series I discussed the empirical, rational, and subjectivist approaches to problem solving. The recent tax debate has highlighted these different approaches and their pitfalls. The Democrats argue that there is no empirical evidence that tax cuts for the rich stimulate the economy. The Republicans make various “rational” arguments that cutting taxes “across the board” will lead to increased spending by everyone (the rich included), and will thus stimulate the economy. Up in Alaska, Sarah Palin takes the extreme Subjectivist approach — a sprightly gung-ho attitude is what this country needs to get us out of the doldrums.
Obama leans towards empiricism. What evidence do we have for taking a particular course? What has worked in the past? In some ways this is a thoughtful and intelligent approach to decision making. In other ways it’s driving forwards while looking out the rear window. Patterns that we perceive in looking at past events may or may not show up in the future. The “empirical fool” thinks “This has happened before, so it will likely happen again.” Well, maybe. But if the system is ruled by chaos and flux, probably not.
The “rational fool,” on the other hand, thinks “It will be different this time, for this and this and this reason.” And thus, a new market bubble is born.
The “subjectivist fool” thinks “I will make it so by imagining it.” This is a great approach for novelists, and in some cases leaders and visionaries. It’s part of why some people find Sarah Palin compelling as a leader, despite the fact that she is ignorant of world events and basic geography. Many liberals thought they were signing on to a subjectivist visionary when they supported Obama (“Yes we can! Sí, se puede!”). What we got was a cautious empiricist (personally, I wanted a cautious empiricist centrist pragmatist in the White House, so I’m not experiencing as much remorse as many liberal idealists).
The subjectivist approach breaks down when it meets intractable, systemic problems, like inoperable brain tumors, or the California budget deficit. Arnold Schwarzenegger, the man who once seemed capable of imagining his way into the office of President of the United States (against all odds, through sheer willpower and charm and visualization), met his match in Sacramento.
The truth is, nobody knows what’s going to happen in the future (and thus, nobody knows the absolutely best course of action for the present). In stable systems, empiricists will have an edge. Rationalists operating on sound premises might have an edge if evidence can’t be collected, or if a system is in a state of flux. Subjectivists might have an edge if a system is at a tipping point, “ready to change” if the right attitudinal forces are applied. A good leader needs to master all three (or more) approaches (and not just while getting elected).
What other approaches exist? How else can we pass through the sticky wickets life puts in our way?
The Intuitive/Super-Conscious Cognition Approach
Sometimes we need to make a decision instantaneously. In some cases we act on instinct. This isn’t really decision making — it’s just reacting. A large dog charges us, and we find ourselves running the other way without even thinking about it. In some cases we might be able to override our instincts — to stand our ground while the animal charges at us. Sometimes our instincts serve us, sometimes they don’t.
Our instincts are reactive programs that evolved to work more often than not. We instinctively feel revulsion at certain scents — but that doesn’t mean the scent itself is universally revolting. A dog, having evolved as a carnivore (acidic saliva, short digestive tract) might find a pile of vomit to have an interesting, or even palatable smell. Humans find the same smell to be revolting and repulsive, because that instinct serves our biological interests. A dog will probably not be injured by eating vomit, but a person might be. We are instinctively drawn towards things that will help us survive and thrive (warmth, human contact, good food), and we are instinctively repulsed or terrified by things that have, in our evolutionary past, injured us (rotting meat, large carnivores, snakes, isolation, even the dark).
Intuition, on the other hand, is something entirely different. Intuition is an instantaneous impulse to act in a certain way based on the sum of our experiences of a certain type of situation. The impulse might also be the result of super-conscious analysis — the kind of thinking our brain engages in when we aren’t paying attention. (I use super-conscious here instead of subconscious because I tend to think of the subconscious mind dealing more with emotional processing rather than analytical problem solving, but maybe the two terms are interchangeable).
An experienced doctor has an intuitive sense about a patient’s condition. A hypochondriac might go to the doctor, terrified that they have a heart condition. An experienced doctor, one who has seen both heartburn and actual heart conditions many times, will probably know instantaneously whether to prescribe antacids or open heart surgery (in the most extreme cases). Of course the doctor would run tests before doing anything dramatic, but their prior experience would inform their “gut feeling” about the situation.
Intuition/super-conscious cognition can work more slowly as well. Our brains are always working — usually on one thing at a time. Paul Graham’s essay The Top Idea In Your Mind provides some great examples. You don’t want to waste your super-conscious processing power on petty disputes or money worries — you want your brain to always be engaged in something interesting and useful.
To engage our intuition, we just need to get out of the way. Instead of actively thinking about a problem, we just make sure our brains have the key facts and then do something else for awhile. Just because we aren’t sounding out words in our heads doesn’t mean our brains aren’t actively engaged. In many cases (Zen Buddhists would say all cases) the best use of our attention is quiet awareness, not active thought. By developing a mental state of quiet awareness, we can capture and utilize productive thoughts (ideas, solutions, intuitive impulses) at their source, before they are lost and forgotten in a constant stream of noisy internal dialogue (aka “monkey mind“).
Sometimes you can access your intuition by taking a different path first. Let’s say you try to make a decision by assigning “Pros” and “Cons” to each choice — a rational approach. Maybe you even assign point values to each item … some Pros and Cons are bigger than others. At the end of the process, one choice clearly outranks the other one in terms of points. But then, you suddenly realize you want to do the other thing. Congratulations — you’ve accessed your intuitive intelligence. Go with your gut.
Malcom Gladwell discusses rapid decision making in his book Blink. He doesn’t like to use the word intuition, because he associates that word with emotional reactions, as opposed to rapid analysis. Still, he’s mostly talking about the same process I’m referring to, except he’s focusing on rapid decision making, while I think intuition can also serve us over longer, slower time scales. One of the conclusions Gladwell comes to is that gathering more information about a situation, and analyzing it for a longer period of time, doesn’t necessarily yield a better decision than the 2-second “gut” decision.
Intuition doesn’t always serve us. Some things that are true are entirely contrary to our previous experiences of the world. Quantum mechanics, for example, is often described as “counter-intuitive.” It just doesn’t fit our experience of the world, in any way. We can’t understand quantum mechanics intuitively, the way we might understand Newtonian physics. Any child who has played tag can apply those experiences of velocity, mass, acceleration and deceleration, to their intuitive understanding of Newtonian equations. Not so with quantum mechanics.
Is there a difference between intuitive decision making and stereotyping? Not necessarily. We can’t really help it — our prior experiences of a group or situation will inform our reactions to all members of that group and all instances of that situation. We’re all prejudiced, one way or another. Our prejudices can sometimes serve us, but if we let ourselves be ruled by prejudice, we will miss unique characteristics of individuals and situations.
The Intuitive/Super-Conscious Cognition approach works best when:
- We have a great deal of personal experience in regards to the problem or situation.
- We need to make a decision instantaneously.
- Our experience is wide and varied enough so that our prejudices will be tempered by numerous counter-examples (broad/cautious prejudice rather than narrow/sure prejudice).
- The situation in not inherently counter-intuitive.
The Holistic/Network Analysis Approach
If your problem involves many different actors — multiple agents with free will and/or distinct agendas and behavioral patterns — then it may be difficult to plot an effective course of action. Who do you approach? What do you ask them to do? Where should you apply pressure? Who should you ignore? What kinds of actions will backfire, or yield zero results?
Right now one of my clients is experiencing elusive technical problems that are due to a combination of an underpowered database resources, possible database administration problems, possible network problems, and code that is not optimized to work over a wide-area network. The last factor is the only one I can directly influence, but in order to offer perceived value to the client I need to fix the problem. It doesn’t matter that many factors are outside of my direct control. My options are 1) to sever the relationship with the client, or to 2) take on a problem that involves about a dozen strong personalities, and just as many less-than-optimal interconnected technological components. I’m opting for #2, for the sole reason that I like many of people that work at this company, and I want to help them if I can.
Network analysis could be considered a subset of rationalism, but in many ways it’s its own beast. In the same way you would approach a standard logic problem, you assign symbolic values to concrete entities. But if you’re analyzing a network of human (or at least biological) actors, then the symbol/object relationship is going to be pretty fuzzy. You can only vaguely represent a complex entity with a symbol, or node point. Network analysis is a good approach if you have no f*cking idea where to put your efforts, and you just need to know where to start pushing.
An important part of network analysis is to determine who, in a given ecology, wields the power and influence. Who makes what kinds of decisions? Who influences the decision maker? If you want to change the outcome, should you appeal directly to the decision maker, or to someone who influences the decider? A child ultimately decides if they’re going to do their homework (or not), but the child’s parents can hugely influence that decision by withholding TV and videogame privileges until the homework is done. That’s why teachers need to touch base with parents now and then.
The ecologist Eric Berlow gives a superb example of how analyzing a complex situation from a holistic/network perspective can illuminate the “pressure points” of the network. He recommends identifying the “sphere of influence” around the node you are trying to change. There are probably many nodes in the network you can ignore altogether, but there are likely a few “key” nodes (almost always more than one) that will require tweaking in order to achieve the desired outcome.
Back to my client’s problem — I decided that to even have a chance at solving the problem I would need:
- a testing environment that mimicked, as closely as possible, the environment in which end-users were experiencing the technical problem
- an adequate budget for troubleshooting and recoding
- full cooperation and open, non-blame-oriented communication from both the database and network administrators
I’ve asked for these things (the first two explicitly, the last one “leading by example”). We’ll see how it goes!
The Holistic/Network Analysis approach works best when:
- There are many individuals or entities involved in the system you are trying to influence.
- There is deep interconnectedness among agents/nodes.
- It’s not obvious where you should be applying your efforts.
I’ll save the last, and potential most powerful, problem solving mode (The Evolutionary/Massively Iterative) for Part III of this post.