Successful habits through smoothly ratcheting targets
Adopting new habits is hard! What a shame: New Year’s resolutions could represent such a bright spark of optimism. Instead, they’re a clichéd punchline on the futility of human will. Certainly, my own past attempts have deserved those jokes!
But in 2017, I shifted strategies and successfully built four new habits (of five attempted): piano practice, internetless mornings, carbless workdays, and meditation. In past years I’d feel lucky if I built just one new habit! I’d like to share my approach: smoothly ratcheted targets, in moving weekly windows, with teeth. Before I unpack that, let’s cover some background.
Most people seem to attempt habits informally, deploying only hope and good intentions. Alas: that doesn’t work very reliably—or at least it certainly doesn’t for me! More serious approaches typically employ one of two formal strategies: chaining, or scheduling.
Chaining is basically “cold turkey” habit adoption: start doing it every day—and don’t break your chain! This approach combines an incentive system (maintaining streaks) with an adoption system (“start doing it every day!”). Streaks can offer a powerful incentive for an established habit, but they do little for fragile new habits. It’s the adoption strategy that really matters initially, yet “start doing it every day!” is a brittle—and, to me, unrealistic—adoption strategy.
Seeing that daily compliance is too ambitious, others try scheduling their habits: let’s play piano on Monday and Thursday nights! In my experience, this works intermittently, but it’s ultimately too brittle in the face of a life I’d want to live. Say Thursday rolls around, and over lunch, a friend invites me to an interesting evening event. No problem: the spirit of my goal would be satisfied if I practiced over the weekend instead. But because I’m depending on explicit scheduling to trigger practice, now I have to actually move a calendar event. That’s high-friction, so instead I just accidentally drop the ball until next Monday. Others tell me this matches their experience.
I’ve found that to reliably adopt a habit, I need a strategy that bends, not breaks, while still holding me accountable over time. It should supply pressure smoothly and flexibly.
So now I start each habit with a low weekly goal: e.g. meditate once per week, any day. That bar is low enough that I don’t have to schedule it or do anything special. Once that stabilizes, I ratchet up the target frequency.
This sounds pretty simple, but I found there are some important subtleties which have made implementation tricky. I’ll outline the key issues now before describing my own solution.
Smooth pressure demands moving windows: make every day doable
Being human, I’d often put off my habits to the end of the week. That worked reasonably well until some goals targeted three days per week. Then it was way too easy to end up in this trap:
To avoid trapping myself, I’d have to perform a sort of time traveling lookahead each day. “Will I accidentally back myself into a corner if I don’t do the habit today?” It felt like my system had pointy edges I had to be careful of.
I’ve found that smooth pressure demands we ensure that every day is doable. There should never be an “impossible” day like the one I illustrated above!
We can make that happen if we use moving weekly windows instead of calendar weeks. Every day, look back a week and make sure you hit your goal within that 7-day window. A moving weekly window prevents us from reaching the situation I illustrated above because on Friday, it would have made clear that we needed a check:
With a moving weekly window, I can tackle habits on whichever days seem best, but I always get a clear warning when today is a “jeopardy” day. I don’t have to look into the future to figure out whether I’ll be able to hit my goal.
Smooth ratcheting and flexibility through fine-grained targets
When it’s time to ratchet up the target, adding one day per week to a habit can feel like a huge change! I find that fine-grained values work better when possible. For my piano practice, I don’t use “numbers of days practiced per week”: I use “number of minutes practiced per week.” I barely notice adding ten minutes per week to the goal, so I can smoothly ratchet up my target.
These fine-grained values also offer more flexibility. Say I want two hours a week of practice. That could be two big weekend practice sessions, or 15–20 minutes per day, or a 1-hour session with a couple smaller sessions. All those configurations hit my target, but the flexibility helps me maintain specific goals in the face of my shifting daily interest and availability.
Give the goals teeth
Accountability is powerful! Certainly, in the workplace, I find I achieve my goals more reliably when I have some real skin in the game. The same appears to be true for personal habits.
There are lots of ways to arrange accountability; I suspect this is probably the least important specific of my approach. It’s probably only important that there be some consequence to slipping, or maybe some incentive for not slipping. Maybe a “don’t break the chain!” incentive would be enough for my habits at this point.
I’ve arranged to lose a small amount of money ($10–$30) when I miss a goal. I only had to pay a couple times in 2017. It’s strange how effective that threat is once a habit is reasonably stable. I’m fortunate that this isn’t an amount of money I’d stress about, but I feel so indignant about losing money for stupid reasons that the threat keeps me on track. I’d probably feel similar motivation from some mechanism which loudly texts my friends when I fail.
Implementing the habit goals
I’m not a “quantified self” adherent, and I’m not interested in graphs or an elaborate dashboard for my habits. But the mechanism I’ve described is complicated enough that a tool really helps implementation. I really only need a small piece of software to tell me “hey! you need to do X habit today!”—and to hold me accountable if I don’t. Occasionally I want to ask a hypothetical like “if I don’t do it today, what would that force the next few days to look like?”
The best tool I’ve found for implementing this strategy is Beeminder, which happily handles all the nuances I’ve described above. Unfortunately, I always have to add an asterisk when I mention it: Beeminder’s great; it’s truly changed my life; yet it’s the highest-friction, least-polished software I regularly use. Most of my friends who’ve tried it can’t get past its interface. This is a shame, but I encourage you, dear reader, to push past the design problems and give it a try anyway.
I’ve looked at many alternatives. Some are very nicely designed! Unfortunately, all are missing at least one of the attributes I described above; in particular, I haven’t found another with the running window, and most lack an accountability mechanism.
So far, I’ve successfully restrained myself from building my own tool. That’s a dangerous yak to shave! But if you find another tool which can implement the behaviors I describe, please do let me know.
I’m about to start a few new habits in 2018, and I’m excited because after last year’s success, I think they might actually stick. Best of luck with your own habits, and happy new year!
When we’re in execution mode, we swim in signs of our progress. A day’s work stands crisply in some endless checklist: four problems solved; three features added; check and check. Am I on track? Did I have a good day? Typical productivity advice applies: set goals; track progress; triage tasks. A checklist makes the day’s grade clear enough.
Working in this mode, it’s easy to feel satisfied at the end of a good day’s work… and (at least for me!) impossible to feel satisfied reflecting on a good year’s work.
Yes, all those tasks got checked off, but if you can easily reduce the work to a checklist, how groundbreaking can it be?
For more open-ended problems, much of the challenge lies in figuring out what to do next. These rich questions offer deep satisfaction on longer time scales, but without a clear sense of progress, each day ends ambiguously. Was today good? Will these tinkerings add up to anything? In what timeframe? Who knows. Ultimately: what structures around progress, self-correction, and operations can help us in open-ended mode?
These questions are intensely personal, but I hope that notes from my journey here may help your own.
Deontology and open-ended work
Creative disciplines from painting to research offer one fairly consistent piece of advice: butt in chair.
This advice usually emphasizes performing actions, since checklist-style productivity advice about achieving objectives applies poorly to open-ended domains. A novelist might begin their day by setting an action-oriented goal to write 1,500 words in a day, or to complete eight uninterrupted half-hour periods. An achievement-oriented goal like “finish writing a chapter,” would be too coarse to be helpful. To break down that goal further would itself require open-ended work.
I started here, journaling my daily time-on-task in pursuit of “deep work.” That structure can become an actionable checklist: e.g. “sketch for three hours from my earlier brainstormed mind map.” Longer focused working periods did help me accomplish more, but I still felt uneasy about the macro-level questions I’d been asking.
For instance, I often felt in hindsight that I’d spent those hours focused on the wrong things—and the vague sense that I could have known as much in advance. More worrying: those journaled hours said nothing about progress towards my true goals.
If I were to repeat my day, should I have sketched for three hours? Should I have sketched those ideas for three hours? Should I have sketched using some other method? Given the progress I’ve been making, should I continue down this path tomorrow or try some other route?
These questions aren’t directly answerable… but that doesn’t mean we should ignore them completely.
There’s an analogue here to deontological ethics, in which actions form one’s moral basis, not consequences. For example, some of deontology’s adherents would argue that a lie is never justifiable, no matter the consequences. Rightness comes from telling the truth because that’s what enacts an individual’s moral imperative.
Action-oriented “butt-in-chair” advice, taken too literally, has a similar bent: a focused creative session is a great outcome, even if all the work has to be thrown out at the end of the day. Success comes from focused work because that’s what enacts an individual’s creative imperative.
Yes, that discarded work often kindles the next day’s brilliant ideas, in the same way that truthfulness in the face of nasty consequences often leads to great outcomes. But that’s not always so; some discarded work is better than others; some sticky situations admit more truth than others.
We’d like to be able to tell the difference. Just as we must seek a more fluid ground in ethics (neither pure deontology nor pure consequentialism), we need more fluid approaches for our open-ended work.
Considering consequences
In open-ended work, goals around a single day’s outputs are misleading or ambiguous, but goals over weeks and months are more concrete.
A graduate student may not know exactly what to do on a given day, but they know that their thesis is due in six months. An artist may not know what to do on a given day, but they might know they’d like to have a gallery show by the end of the year.
Medium-term achievement-oriented goals can support error correction and provide feedback to action-oriented day-by-day goals.
For instance, say that I’ve been checking off “sketch for three hours”-type tasks every day for two weeks. I can reflect on the progress I’ve made in hindsight and compare it to my hopes for the following month. If I’d hoped to try a prototype in classrooms before the school year ends, and my sketches aren’t yet converging on a single idea, I might shift the focus of my creative work to depth over breadth. On the next day, I might choose my five favorite concepts, then set action-oriented goals around making them higher-fidelity.
If I’ve not generated many ideas despite endless diligent hours in the chair, I might shake up my methods, limiting myself to ten minutes per idea, or only allowing straight lines, or trying a different medium. If that doesn’t work, it might indicate I’ve reached an idea cul-de-sac. Time to backtrack.
This kind of planning-oriented thought can crush a more expansive, ideating mindset.
In the middle of my sketching hours, I don’t want to be worrying about whether I’ll be ready for my classroom prototype next month. Within a given day, action-oriented “butt-in-chair”-style advice does help; meta-thought is just distracting. But go too long without error correction, and you’ll misspend hours in the chair. Some separation is in order.
My current practice
I’ll outline my current approach now. It’s young and evolving, but I certainly wish I’d been able to read this a few years ago, so perhaps it will help some readers.
I begin each day by selecting some action-oriented goals which I hope will advance some broader achievement-oriented goal. For example, if I’m working towards an in-classroom prototype around a set of ideas, I might aim to spend three focused hours fleshing those ideas out in sketches.
This is a natural spot for brief deliberation, but once the day begins, I focus on the actions I’ve chosen and suppress planning. The rest of the day’s work becomes roughly deontological. I give myself permission to be satisfied with the day if I spent three focused hours sketching like I’d planned.
Weekly, I reflect on the previous week’s mix of actions and my progress towards the broader goal. I consider what’s working and what’s not, then I make a few notes about how I’d like to adjust my daily mix of actions in the next week.
I always assign some target date to that broader goal (“run a prototype in classrooms by June 1”). My weekly reflection pushes on that date: does it need to move? Does my daily work need to change focus to hit it? These higher-level reflections help me feel a sense of progress (or note a lack of one), and their regular course corrections give me the safety to draw satisfaction day-to-day from ambiguous, action-oriented goals.
Monthly, and when completing a broader achievement-oriented goal, I reflect on the bigger picture which defines my goals. That roadmap invariably evolves as my work proceeds, but maintaining that long-term plan helps me connect my present work to downstream goals. Those connections themselves inspire satisfaction: they help me see a path to my true goals.
For this level of reflection, I’ve found it helpful to reference others’ related roadmaps. For example, here’s IDEO’s product design process; my design partner May-Li Khoe uses a similar process of her own devising. Every field has dozens of competing structures like this, but their underlying similarities can provide useful scaffolding.
From time to time, I flip back into execution mode. It feels like an old friend. We say hello, dance for a while, and part ways smiling, just as it always was.
Open-ended mode is more enigmatic, reserved—yet occasionally it sparks some moment so singular it lights up the whole year. Those moments don’t happen without the days spent together between those moments. I’m slowly learning to make the most of our quiet strolls.
Science developed out of philosophy. Before we had Copernicus’s planetary tables or Newton’s equations of motion, we had Aristotle’s rhetoric. That medieval natural philosophy was wrong, but it still made useful predictions.
I’ve spent my career making things in the worlds of consumer products and education, and in both domains, I’ve observed a common failure mode in decision-making: an overriding obsession with data—with appearing scientific—and an associated repudiation of philosophy.
That’s a totally appropriate obsession in some fields, like manufacturing, transportation, and aviation. But in consumer products, education, and so many other domains involving the messiness of humanity, the data obsession falls prey to hidden errors and distorts our true goals. Worse, it deprives us of truly meaningful insights that are available via philosophy, intuition, and stories, but not yet fully explicable through quantitative systems.
This danger lurks in domains that have not yet been systematized. When it comes to people, we lack Newton’s equations of motion. Actually: we don’t even know what the equivalent equation should be measuring. Even if we did, we probably don’t have instruments that can measure those quantities. Can there ever be equations which can usefully describe these phenomena? I think so, but I don’t know we can even be sure of that.
Until we build more powerful explanatory theories of these domains, we must respect the role of philosophy and beware the dangers of playing scientist.
Measuring everything, measuring nothing
Let’s talk about test scores.
Have you ever gotten a fine grade in a class—say, calculus—but later felt like you didn’t really understand what was going on? That you can follow the steps you’d learned to solve problems like ones your class tackled, but you can’t explain why they worked or apply them in new contexts? This experience seems totally ubiquitous!
So: how much do you trust test scores for making decisions about educational systems?
In fields like education and design, we measure only indirect proxies: page clicks, time on site, changes in test scores, survey responses, and so on. Then we try to make decisions with those measurements.
This is like wiggling a rod, attached to a complex set of gears, attached to the thing we want to measure, which in turn is attached (in mysterious ways) to the thing we’re actually measuring.
When we don’t truly understand the mapping between what we really want to know and those proxies, we easily miss important consequences.
Hours of practice worksheets at Kumon might make a child great at arithmetic, but what does it do to their curiosity? To their desire to learn independently as an adult?
When Spotify opts you in by default to noisy push notifications (“the Beatles are now on Spotify!”), they might increase some engagement score, but they also annoy their users. That annoyance may not show up in any dashboard: maybe users keep using the service exactly as much, but when some PR fiasco blows up the following year, they’re less inclined to take Spotify’s side.
Because we don’t understand this mapping, we have to make many more guesses. And with every guess, there’s some chance that we see some result purely by chance. Statistical hypothesis testing only has meaning when you account for all the hypotheses you’ve tried.
Alternately, sometimes people don’t even make guesses, and they just go hunting in the data. If you go looking for patterns in a sufficiently large data set, you’ll certainly find some!
(”Significant” from XKCD by Randall Munroe)
We know that correlation doesn’t imply causation. Sometimes, you’ll find strong correlations by random chance—just like in the comic above: the data suggest that this shade of blue causes people to engage the most!
But it was just random noise, and if you recolor more elements in that shade of blue, you won’t actually make anyone happier, other than perhaps your own community's navel-gazers.
Another danger is that your correlations may be masking more important underlying phenomena.
Say you want people to share their big news on your social network first. You can’t measure that directly, but you have a proxy metric: photos included with those posts have EXIF data indicating when they were taken. You decide you want to minimize the time between the photo being taken and the photo being shared.
To figure out how to proceed, you hunt for correlations in your logs of users’ behavior. Say you discover a strong correlation between how quickly a photo uploads and how likely it is that users share photos of big news immediately. You tell your engineers to focus on optimizing upload time!
You ship your optimized photo uploader… but you don’t see any benefits in the metric you were measuring. Turns out, you didn’t see this correlation by chance: you saw it because people with faster upload times can afford better cellular connections, which means they’re more likely to upload photos while they’re out and about, as opposed to waiting until they get to unmetered WiFi.
Photo upload time was itself just a proxy measure of the true root cause.
Even if we’re pretty sure we don’t have any hidden causes or consequences lurking, and we carefully account for all our hypotheses, we must remember that these are proxies we’re optimizing. As the situation varies, these proxies’ connection to your true goals may taper—or reverse!
Vitamin C can prevent disease, when taken in small quantities, if your diet already didn’t contain much of it. But that doesn’t mean you should take a hundred times as much seeking a hundred times the benefit (as double-Nobel-laureate Linus Pauling did): you'll see no marginal benefit, and you’ll just excrete it all.
In the worst cases, fixation on these proxies can create perverse incentives. Say that you want to prepare students for a life of solving challenging problems. It’s true that minimizing missed days of school may help make that happen—but past a certain point, other factors will dominate.
If you optimize too aggressively for zero missed days of school, you may easily reverse the correlation, disrupting students’ family lives or creating an atmosphere which makes students resent their autocratic school.
If you make a product, total usage time might seem like a good proxy for customer joy. But if you take that metric too seriously, you’d be punished for making a change which would help a customer accomplish a given task in less time than before.
There’s a subtler issue with exalting data in these domains—one my research partner May-Li Khoe has patiently explained for me over and over again. If you try to design something with human meaning by steering toward maximum impact on business outcomes, you’re very likely to end up with little human meaning… which in turn will likely harm whatever business outcomes you’re measuring in the long term.
Similarly, “teaching to the test” sucks the fascination and participation out of classrooms in exactly the way you would expect.
This talk from Frank Lantz covers the issue wonderfully in game design (this quote at 33:30; my thanks to Bret Victor for the pointer):
The dilemma of quantitative, data-driven game design.... So here's an analogy: Imagine you have a friend who has trouble forming relationships… "I don't know what I'm doing wrong. I go on a date, and I bring a thermometer so I can measure their skin temperature. I bring calipers so I can measure their pupil, to see when it's expanding and contracting..." The point is, it doesn't even matter if these are the correct things to measure to predict someone's sexual arousal. If you bring a thermometer and calipers with you on a date, you're not going to be having sex...
So.
Imagine that two teachers have exactly the same measured impact on their class’s test scores. How likely is it that they have the same impact on creating empowered thinkers?
You decide to adjust some variable because in the past, it correlated highly with increased product usage. How likely is it that this change better solves a meaningful problem for the user?
Meaning without measures
We’ve seen that there are plenty of dangers in making decisions based primarily on indirect measures with hazy connections to our true goals. Yet clearly, great teachers and great designers do operate effectively in these unsystematized fields!
They have insight; they have intuition. These come from an internalized philosophy about the field, drawn from experience, observation, and stories. Yes, their philosophies are imperfect; and no, they can’t necessarily give you a set of calipers you can use to make your own decisions.
But if you ask about one student interaction in particular, or one product detail in particular, they can often explain in hindsight why their philosophy pushed them in one direction or another. Listen enough, and you might build some intuition of your own.
This is not just luck or some kind of confirmation bias—there’s an underlying consistency to these experts’ taste. It’s clearly visible even if neither you nor they can quantitatively describe how they’re doing what they’re doing. Great teachers sure do manage to consistently be great teachers, in a way others consistently recognize, even without a dashboard and A/B tests. Of course, we might have to watch for a while to see that an expert delivers insights consistently, as opposed to by chance—that’s why knowledge worker interviews are so hard!—but it’s clear that some experts’ ideas are more consistently successful than others.
That consistency is what meaning is.
How do you know your house exists? After all, you don’t experience it directly: your contact with it is mediated by all kinds of layers of fuzzy visual processing and your own faulty memory. It exists because it’s reliably where it was last time. It exists because when you’re inside it, you consistently see the same imagery, with shadow angles mediated as you expect by the seasons. It exists because others can talk to you about your house and say things, interpreted through a winding auditory system, that somehow match your own fuzzy perceptions. It exists because your fingers can feel the shape of the house number on your door, which matches the shape on that lease you remember signing long ago.
The same logic tells us that when an expert consistently makes decisions broadly regarded as successful, and can explain their philosophy with rhetoric that makes intuitive sense, there is probably a there there.
Your house is more systematized—we can precisely measure its height, draw blueprints, predict its mass—but society could still effectively talk about houses before we had any of those tools. Until we discover those tools (and the questions we want to ask with them!), all we’ve got is tradition, expertise, rhetoric, philosophy. If we listen with balanced skepticism and curiosity, those can be powerful tools themselves.
Pretending to measure meaning
I don’t need to preach so strongly. In practice, we usually can’t ignore domain philosophy and experts’ intuition anyway.
Meaningful philosophy is meaningful—so even if we say we’re throwing it out, our intuition often remains entangled with our decisions.
I see this all the time in product decisions. For instance, someone might believe that sign-up walls make for a bad product for a variety of philosophical reasons, but they justify this decision outwardly by pointing to some data from one product’s blog about their A/B test on the subject.
That data is not the reason they decided to ditch sign-up walls. It’s just the reason they’re giving to others (and often, themselves) about why they made the decision. This behavior represents a sort of homage to science… while simultaneously violating its core principles.
In the education space, people are very excited about growth mindset interventions. The rough idea: if you can persuade a child that intelligence can grow with practice and hard work (just like their muscles), then they’ll actually perform better in school.
The recent enthusiasm for interventions in this space follows a series of randomized controlled trials documented in studies by Carol Dweck and her team at Stanford. These interventions probably are effective! But: the field’s quantitative results are actually quite modest in effect size.
The studies alone can’t justify the magnitude of the excitement about this topic; that follows the magnitude of people’s pre-existing intuitive beliefs in these interventions. The problem is that when the education community talks about this topic, it primarily justifies growth mindset interventions with these studies.
This type of motivated reasoning corrupts the dialog around decision-making. We should use provisional data like this to support—not supplant—our philosophies.
When two people disagree philosophically about an issue in an unsystematized domain, but allow only quantitative arguments, they end up fighting a proxy war through data weaker than their own beliefs. Worse: if we ever do invent powerful predictive systems for these fields, we’ll need our scientific wits about us, unsullied by post-hoc spin.
I hope it’s clear that I’m not arguing for us to generally abandon data and systematic thought. This scientistic obsession is a reasonable defense mechanism! After all, before precise measurement, physicists used to debate with rhetoric, and we ended up with phlogiston (i.e. things burn because they contain an element called phlogiston; phlogiston is lost to the air when a thing burns; things can’t burn inside a jar because that air can’t absorb any more phlogiston).
It’s in fields without reliable systems that we can’t measure our way to understanding.
Building systems in those fields is a critical project, and progress can be made. Meta-analysis and multitrait-multimethod tests have certainly helped us lay some foundations. Yet while fields’ systems are under construction, we must be careful not to put too much weight on them. They’re not yet structurally sound.
Intuition, philosophy, and expertise deliver all kinds of useful tentative explanations. If we monitor their predictions over time, we’ll discover limitations, and our theories will evolve. All the while, we’ll spot patterns, incorporate provisional systematic concepts, and fluidly evolve our beliefs, taking the best evidence however it comes.
Joy, belonging, and empowerment may live in this figure’s “qualitative black box,” but we can still produce explanations for how they arise. Those explanations may well involve measurable inputs and outputs. But if we insist on explaining joy through, say, engagement time and Net Promoter Scores, we’ll get exactly as much joy as we deserve.
Imagine that you’ve just moved to a new city. Whenever you need to go somewhere, you ask a special machine how to get there. But it doesn’t show you a map: it tells you to take twenty steps, turn left, go forward a hundred steps, etc. When you step outside, you’re blindfolded, so you can’t observe street names or other landmarks, but after years of traveling this way, you always arrive safely and on time, so you’re not complaining.
Navigating that way, you’ll have no idea how to get somewhere new without those step-by-step instructions. You’re unlikely to discover anything surprising while in transit, you won’t devise shortcuts to your destination, and you won’t notice when one location abuts one you’ve already visited. But this is exactly how we are typically taught!
This approach to knowledge is almost ubiquitous: math students memorize the steps to integrate equations that look a certain way, physics students memorize equations of rotational motion separately from equations of linear motion, computer scientists memorize the properties of various data structures. They know how but don’t understand why. It’s the only way of learning that many people experience throughout their education, so it’s the one they keep applying after graduation.
A student who learns the results of a field—without understanding the explanations which led to those results—is stuck making essentially vocational contributions. He can apply the hows he's already learned in the scenarios preidentified by his teachers, but without the whys used to make those hows, he can’t readily synthesize new solutions or cope with new scenarios, just like our blindfolded city navigator.
Of course, it’s okay not to know everything all the time! Learning why certainly takes quite a lot longer than learning how. You might even motivate yourself to learn fuller explanations by first learning their exciting results. Danger only appears when pragmatic knowledge of hows is mistaken for true understanding—that’s when the knowledge perpetuates its own limitations and poisons surrounding endeavors. I’ve watched it pacify with intimidation, confine with fear, and (most dangerously of all) promote indiscriminate stasis. I see veterans of industry walking around with eyes closed, counting their steps.
I’m fortunate to live and work in a community of brilliant people. It’s certainly important to recognize peers’ contributions and skills, but sometimes I hear myself and others singing a different kind of praise. Breathless praise, helpless praise, praise almost sighing in frustration: “Wow. That guy’s just so damn good.”
Admiration can be productive: it might help distill that role model’s shine into some useful goal (“looks like I need to brush up on my statistics”) or observation (“maybe that approach would be useful in this other problem”). But that “so damn good” praise hides intimidation; it can't offer insight because the speaker can’t even begin to imagine how the exemplar’s doing what he’s doing. It’s like yearning for some destination which our directions machine doesn’t know about, without being able to take off the blindfold.
Knowledge can help us plan a path from point A to point B—from the admirer to the role model. But the directions machine’s step-counting knowledge is too narrow to plan a path to somewhere new. For that, you’ll need a broader view—a map. When I catch myself spouting intimidated praise, I take off my blindfold, start exploring the surrounding neighborhoods, draw a map, work my way back home.
For much of my life, though, I couldn’t lift that veil; and worse, I didn’t realize there was anything to lift. I mistook the step-counting directions I was learning for understanding. I thought how was as good as why. Then those memorized facts and procedures became false knowledge, their presence more harmful than nothing at all.
The idea of acquiring knowledge a different way had never even occurred to me: after all, everyone around me seemed to be getting their directions the same way. I’d never seen anyone else take off their blindfold—how would I even go about it? Who would give me directions to make my own directions? How can I know what I know?
Sometimes this brand of false knowledge induces more active fear. You’re invited to a meeting at some distant location, but your machine has never heard of it, so you have no idea how to get there. Questions build as the appointed time approaches: you’ve always been able to get where you needed to go using the machine. Is it possible that there are some places the machine doesn’t know about? How many places? Are they important? You think of yourself as a pretty good navigator—so are you, really? Has it all just been luck so far?
This is a scary line of reasoning. Each step amplifies the mounting cognitive dissonance. At this point it’s clear that a different approach to knowledge is necessary, but there’s so much pressure to resolve the dissonance that it’s easiest to actively reject the goals that incited all those painful questions.
The alternative doesn’t look great. It means summoning the humility to admit that your knowledge is far more limited than you’d imagined. You need to figure out the whys behind the hows you’ve been using, so that you can make some new hows. For a while, since you’re drawing the map (and figuring out that you need a map, then what a map is), you'll need much more time to get anywhere. The easy conclusion: maybe you don’t need to go to that meeting after all; maybe only “really smart” people understand math proofs; maybe you’ll “let the experts” take care of that tough problem that came up.
Now the situation is much worse than the impotent haze of simple intimidation. The pressure from cognitive dissonance caused active rejection of the path to real understanding.
I’ve seen an even more dangerous hazard emerge when this kind of fear builds, especially after many years of success. Sometimes the mounting cognitive dissonance can’t be fully resolved just by avoiding destinations unknown to the directions machine. You might have to persuade others not to visit those places. You might have to convince yourself that the people visiting those places are foolish or harmful, that their machine will surely lead them astray. Now you’re a zealot.
Of course, you can’t see that those renegades aren’t walking around with blindfolds outside. They’re navigating with a map. Each person has drawn his own, and they’ve helped each other fill in the details.
The convictions I’ve described become dangerous when they fester in someone with power. Overwhelming dissonance will pressure the zealot to yoke his power to perpetuate stasis, forcing others to avoid solutions outside that narrow set he can access, twisting problems into ones which can be solved with answers he’s already memorized. Radical thinking is a threat to the zealot's psychic well-being.
So when this zealot can influence others or controls something that others rely on, those folks are in trouble. New problems will always emerge in any endeavor, and no machine can be programmed with all the solutions up front. Eventually, some problem will appear with no predetermined solution. The zealot will have to muster all his understanding to synthesize a new answer, but he doesn’t understand anything: he’s just memorized a bunch of results.
Software engineering is full of zealots who fiercely perpetuate this kind of stasis: they only know one language, only want to know one language; it’s the best, and all the others are stupid. Steve Yegge writes about the experience of taking off that particular blindfold and learning some of the related whys:
After programmers spend years memorizing the proper incantations, telling them that their hard-won language is going away is like pointing a gun at their family. […] The virus of programming-language religion has a simple cure: you just write a compiler. […] For weeks afterwards, you can't look at your code without seeing right through it, with exactly the same sensation you get when you stare long enough at a random-dot stereogram. […]
It's fun to be a shaman, knowing that typing the right conjuration will invoke the gods of the machine and produce what you hope is the right computation. Writing a compiler slays the deities, after which you can no longer work true magic. But what you lose in excitement, you gain in power: power over languages and over language-related tools. You'll be able to master new languages rapidly and fearlessly. You may lose your blind faith, but you gain insight into the hauntingly beautiful machinery of your programs.
My own story actually started with physics. I spent years memorizing sheets of equations for physics classes without understanding how they were derived, how they were connected, or their significance. I was learning how to compute answers to some common physics problems without understanding any of the whys behind those hows. That’s just how we were taught, and it appeared to be the level at which my teachers understood the material.
I tried to make games with physics-based behaviors and had all the problems you’d expect: I had no idea what to do when confronted with new problems; I was helplessly intimidated by successful game authors; with that fear ever building, I actively rejected the idea of building real understanding, convincing myself that I could get by copying and pasting together code from random internet searches. Thankfully, I wasn’t in a position to put up a real fight for stasis.
In my freshman year at Caltech, one lecture totally reversed my approach to learning about physics. It was the first time I’d taken off the blindfold in any subject. Over the next two years, it led to the painful realization that I had very little real understanding of almost anything else, including computer science—my purported specialty.
The lecture: we’d just finished studying special relativity and electricity. The professor walked in and drew a new line between those two topics. “Suppose we have an electron moving along a wire at relativistic speeds…” He started by writing the familiar equations from electricity and relativity. A few blackboards of algebra later, the equation of a magnetic field just… appeared.
I’d memorized that equation years earlier, but I didn’t know what it meant. I had no idea about the underlying connections between these phenomena. The difference between magnetism and electricity is—literally—a matter of perspective. Something clicked in my brain. That was when I started to open my eyes and ask why.
I recently spoke with some kids from a high school with an impressive engineering curriculum. Classes in mechanical, electrical, biological, and software engineering—the mind boggles! I told them my (enormously contrasting) story: how my high school had no engineering curriculum at all, a math and science curriculum by name only; how I couldn't find any adult engineering mentors in my area; how I had to learn it all myself from books and the internet.
At this point, I think they expected me to launch into some kind of school spirit speech: "and you all are so lucky; cherish what you have here, and when it seems boring, just remember that some people would kill to have these resources; etc." Please don't misunderstand me—these kids are lucky!—but I had a different message in mind.
I pointed out their wonderful bio-engineering class and earned raised eyebrows when I noted that they were learning ideas in that class which had only been known for a few years. I asked the kids how they thought their teachers learned those ideas if not from their own schools.
I told them about how much my own field has changed since I started learning it, how even if I had engineering classes in my high school, many of the skills I'd have acquired would now be irrelevant. Worse—how I'd watched exactly that irrelevance befall peers who'd learned their trade then felt they were finished learning once they'd graduated or added "senior" or "chief" to their job titles.
But, I told them, there's a way out: the long-term defense against obsolescence in the face of progress is to never stop learning.
A software engineer friend of mine is about to finish his undergraduate education, but he told me he planned to get a master's degree. I was surprised to learn that: he seemed to enjoy his internships far more than his time in school. Why would he choose to stay in school?
He explained that he didn't really want to stay in school, but he felt like he needed to: my friend suspected that all good software engineers would need to know machine learning in a few years. He wanted to be competitive with them, so he planned to get a master's degree specializing in machine learning.
I was sure the program he liked would teach him effectively, but I asked him why he didn't think he could teach himself that subject instead and save the money. His confession: he feared he wouldn't have the will-power to make himself study in the "real world."
That's a reasonable fear! But I followed up: what would happen when, in fifteen years, a new subfield emerges which all good software engineers must understand? Would he quit his job and get another master's degree?
My conversation with those kids about bio-engineering continued from that last point.
They'd concluded that their teachers must have learned the subject from books. Fine, so: "Where did the books' authors learn the subject?" Some consternation now! (I swear I maintained my composure when one girl exclaimed "from God!") The kids finally landed on "they learned it from themselves!"
Yes! But why stop with a defense against obsolescence in the face of progress? After all, the best defense is a good offense.
That obsolescence-causing progress has many faces, really. It's not just the researchers and academics who automate away your job or change society's expectations: what about the progress in our consumption of natural resources? The progress of time obsoletes your knowledge by physically destroying it! For now, anyway.
The obsolescence-causing progress created by humans feeds our defense against the obsolescence-causing progress created by nature, but right now, we're outnumbered. The overwhelming majority of our population essentially stops learning at age twenty. In my mind, that's the first thing we must learn to change.
When I was a kid, I spent much of my time making video games. Then I devoted my early teenaged afternoons to making an app for making art for video games. Now I make tools which can (among other things) help make apps for making art for video games.
My goals have shifted away from games over the years, but I still consider the same metaphor that excited me in eighth grade: that of leverage. It's a useful principle no matter one's goal, so I'd like to share a few examples.
Say you want the world to have amazing teddy bears. You could make them yourself, but you couldn't make enough for everyone! If you spent your time teaching people how to make especially wonderful teddy bears—and if you did a very good job—the world might have more amazing teddy bears in the end. Your effort is effectively magnified, leveraged. You might also amplify your work by researching extra-fluffy stuffing and selling the results to teddy bear manufacturers.
So, no more video games: now the point of my work is to expand the reach of human knowledge. Computers and software are outrageously useful implements for that end. We've got a world of knowledge at our fingertips, along with the tools to explore, share, and expand it. Those tools give us leverage over knowledge. An hour at a library can become seconds with software like Wolfram|Alpha.
Presently, I make tools for making tools. On the occasional days when I do something that radically improves the quality and ease of creation of all these levers, the effect compounds into fantastically broad reach.
I've long mused that a teacher could have much greater leverage over the reach of human knowledge: after all, he trains the people who could go on to make tools for making tools for creating knowledge! Among all the other wonderful things his students might do.
In the last couple years, though, I've enjoyed considering the principle taken one step further: what if I created tools which allowed teachers everywhere to be radically more effective and efficient? Or which helped students teach themselves far more easily? Leverage abounds.
Need I remind you: the human race now sends robots to Mars. That is a thing we do! We derive that absurd capability from generations of previous epistemological revolutions which we compound and extend to create new marvels.
I like to picture the tower of abstractions that enable this kind of progress. If you looked closely, you'd see its struts form an intimidating (and not altogether reassuring) lattice. Rocketry builds on chemical engineering and relativity but not pharmacology; flu vaccines build on chemical engineering and pharmacology but not (directly) relativity.
Now, we could examine the same data or questions from any position on this tower. One fun example: why is your hair the color that it is? Because you have a certain gene which expresses that phenotype! Or… because of the cellular machinery which expressed those genetics? The electrochemistry which drove those interactions? Optics? Cultural tropes influencing the mating habits which favored those genes?
One approach suggests that the observer can best understand a complex system by breaking it into composite parts, until she trips upon axioms, which are (perhaps temporarily) where truth lies. This is sometimes called reductionism. This word has other—incompatible!—definitions, but this is the one I'll reference.
Meanwhile, others suggest that for many complex systems, the constituents hold only noise: real meaning emerges at the highest levels. That interpretation seems especially appealing when considering human consciousness and free will. After all, who wants to think of their being in terms of roiling, deterministic chemistry? I'll call this approach holism (with the same lexicological caviats).
This is a figure from the "Ant Fugue" in Douglas Hofstadter's Gödel, Escher, Bach. It contains letters made of other letters. At the highest level we read "MU." The 'M' comprises "HOLISM" three times; the 'U,' "REDUCTIONISM." Each letter of the former contains the latter word, and vice versa. Then, gleefully (and invisibly in this reduction), the smallest letters are themselves formed by repeated "MU"s.
Each level of abstraction in this figure encodes information (here, words) in its emergent properties. If you näively apply reductionism and interpret the smallest components, or take ten steps back as holism prescribes, you read only "MU." In Chinese, "mu" means nothingness, often in a spiritual sense.
It's not impossible to discern the middle layers of words from either of these positions, but it's also not expedient! Independent access to each layer of abstraction delivers novel perspectives and understanding. I suspect the same is true for any complex inquiry.
Now then: how does one acquire that independent access to each layer? Which reduces to asking how one can build expertise in general—how can we learn the many facets of any complex field?
A key issue in education is that the student's goals are often holistic. She wants to learn how to make a robot, for instance, and might not be interested in the complex analysis necesary to understand the physics necessary to understand the electrical engineering necessary to understand the servos and sensors necessary to make it move.
Two dangers lurk here. The first and more apparent: if the student plays only with pre-built motors and components (without understanding or considering the principles by which they're constructed), she'll be seriously limited in how she can compose them and will endure endless headaches diagnosing and fixing problems.
The second and more insidious: this naïvely holistic (vocational?) approach could appear to work for quite some time. A year or two later, when she abuts the edges of the epistemological landscape visible from her position, she'll face enormous psychological barriers as she moves down the stack. She must unlearn parochialisms, summon humility in a field she thought she understood, muster patience for moving—temporarily!—more slowly relative to her ostensible goals.
But if she builds her tower of understanding from principles—brick by reductionist brick—boredom may dominate good intentions and drown her emerging motivation.
I propose a hybrid approach: derive motivation and intuition from top-down exploration while cementing understanding and that crucial universal access with a bottom-up climb. Critically, though, she can ground and inspire the lower levels by studying them in context of her experiences with their emergent behaviors.
Some schools will offer a physics class for physics majors and a physics class for engineers. But generally the latter distinguishes itself by watering down the concepts, skipping a great deal of material, and hand-waving over derivations—not by providing motivating context.
With this idea in mind, personalized education could provide enormous benefits not just in terms of keeping pace with the student's progress and prior experience, but also in captivating him through critical foundational material which would otherwise never take root.
I'll leave you with a personal anecdote. At Caltech, I spent two years studying not computer science (my major) but math, physics, chemistry, biology, astronomy, laboratory practice, etc—because I had to! Every student had the same core requirements, regardless of major: Caltech history majors have studied statistical thermodynamics.
At the time, I was incredibly bitter about this situation! I felt like I was wasting my time, so I didn't take the material seriously. I was supposed to have learned linear algebra my freshman year, but I didn't: I learned it after I bumped up against a brick wall in my studies of machine learning, which I'd practiced for a few years with misleading success.
So I went back to re-learn linear algebra in a real context. This time I cared, so this time it stuck. This time I understood how these abstractions might be applied, so they made sense. Which was nice, because years later they enabled me to (more or less on a whim) radically improve how iOS's 3D "page curl" effect tracks a user's finger.
I've had similar experiences with those physics, chemistry, and biology classes. They're a means, not an end, but it wasn't until the end was in sight that I could truly learn and build from their subject matter.
Some folks have ethical objections to eating meat, but no one seems to mind savaging a carrot. Why? Both fish and ficus are living organisms, after all. I doubt vegetarians consider taxonomic kingdoms when making moral judgments: the decision comes more intuitively.
Douglas Hofstadter suggests that many are judging the presence of consciousness, a sense of self-hood, "inner light." He argues that this attribute isn't just an on–off switch but rather a continuum: humans seem to have a stronger sense of "I" than a dog than a fish than a fly. Your sense of species' relative positions might be different from mine—I'll dodge the question of what specifically we're assessing—but surely you'd place, say, monkeys and ferns in different spots.
Now let's zoom in on the "human" portion of that spectrum. Does a newborn occupy exactly the same spot as you? Infants don't recognize themselves in the mirror, don't appear to reflect upon their experiences, can't plan, couldn't pass the Turing test. They're closer to you than a fish, sure, but how do they compare to a chimp?
Someone recently asked me if I'd ever been to Colorado. I said "yes, but I was nine-ish. Not sentient yet." My reply was flippant, but it reflects some truth: my "inner life" at that age was far simpler than the one I've enjoyed in adulthood. My behavior was almost purely reactive; I integrated facts without understanding; I exhibited limited theory-building or subjunctive thought. I was closer to an "i" than to an "I."
That made sense: consciousness seems to be a feedback system constantly evolving in response to perception, so it should grow in reach as it consumes more input.
But does our consciousness grow monotonically? Consider the last time you were quite drunk—after all, alcohol is beloved in part because it makes us less self-conscious! Other drugs make their users more self-conscious, while still others move one not exactly towards "I" nor "i", but sideways into surprising alternative modes of awareness.
So perhaps consciousness grows monotonically when not adulterated by external chemical influences. But then what about internal chemical influences? How did your sense of self respond the last time you stayed up all night? I find (to my great alarm) that if I lose even an hour of sleep, most factors I associate with my consciousness suffer hugely!
Conversely, if I find myself unable to perceive, reflect, analogize, or emote, then a peaceful walk often restores these faculties somewhat.
This is all fairly terrifying: our incredibly intimate sense of self ebbs and flows by the hour with unseen tides. But if it's going to happen, perhaps at least we can exert some control over it! Can we exploit the "peaceful walk" mechanism not only to restore these sentient faculties but to boost them? Longer-term, could certain hobbies or lifestyles lead to a more powerful sense of self? Which, while we're introspecting, demands a further meta-question: would that even be desirable or useful?
People list plenty of reasons why kids should learn how computers work. They could automate repetitive tasks; they'd be empowered to create in all kinds of media; they'd learn powerful new problem solving approaches. That last point (so innocent seeming!) has gripped my imagination most thoroughly: a mastery of abstraction offers stupendous power in practically every endeavor.
But perhaps the most important tower of abstraction—one whose structure we scarcely understand despite its supremely personal importance—is rarely mentioned in this context: life.
Say we cure cancer. Hooray for us! But we'll face other diseases, then still more after we've addressed those. We address the analogues in software (bugs, security holes) at a totally different pace from those of life. Even the attitude is different. Software engineers know their software has bugs—and that it probably always will—but the industry nevertheless accelerates and thrives because when issues crop up, they're generally resolved quickly and decisively.
We can fix software bugs quickly because we (mostly) understand the systems we're addressing at all relevant levels of emergent behavior. We've established that understanding through composable abstractions about which we can reason in isolation, and by employing tools which enable us to rapidly test conjectures and gather data.
Personalized medicine applies some of these principles to healthcare: when a patient's bacterial population evolves a resistance to a drug, we should be able to rapidly study that resistance and produce an appropriately targeted treatment. For this end, we'll need diagnostic and fabrication tools, yes; but more fundamentally, we'll need medical understanding at the many levels of emergent phenomena between organic chemistry and the consequent symptoms we observe.
Software engineering could provide a powerful microworld for biological engineering. A generation of children steeped as much in abstraction as in language would become a generation of adults extraordinarily well-equipped to understand our biology—and therefore to manipulate it. It's difficult to comprehend the reach of explosive growth in this field. How many more generations will die of old age?
Abstractions empower and accelerate. As usefully encapsulated nuggets of understanding, the creation of novel abstractions drives a field's progress, but their invention is possible only with deep understanding of present ideas. So I declare: if we are to master a field, we must accept none of its abstractions as magic. Rather, we should yoke them as automations of what we already understand.
This concept separates the random, slowly-reinforced walk of genetic evolution from the directed, rapid exploration of scientific evolution. Understanding provides not only a compass bearing but also course correction, without which extended progress is impossible.
For instance, if you always cook from recipes (without understanding why you're executing each step), you'll be poorly equipped to fix a dish that's gone wrong, improve an average one, or synthesize something new altogether. The same limits apply to parochial understandings of abstractions in musicianship ("minor chords sound sad"), writing (the five-paragraph essay), mathematics (the Pythagorean theorem), and engineering (caching systems).
Say you accept this idea. We're then faced with a practical problem: how can you ever learn to do anything useful without endless prerequisites? Algebra classes are particularly afflicted.
TEACHER: Today we'll learn how to factor quadratics of this form.
STUDENT: Why do I need to learn that?
TEACHER: You could use it to find where parabolas hit the x axis.
STUDENT: Why do I need to learn that?
TEACHER: You could use that to find the trajectories of point masses.
STUDENT: What?
TEACHER: Believe me, you'll be glad you learned this someday.
To foster real understanding, we must format concepts so that students can derive solutions for themselves, not encourage them memorize some procedure. And of course, they must be motivated to find those solutions by a problem they actually care to solve.
Microworlds may form the basis of one solution. The idea is to create a tiny, self-consistent sandbox in which the student can explore some concept, given little instruction beyond what he already knows. The sandbox doesn't hide detail so much as focus. One can imagine, then, constructing a successively expanding universe of these microworlds—each itself useful—so that hops between neighbors aren't too vast.
Imagine trying to learn to play the keyboard. You could start with one which only allows the student to play in the pentatonic scale in C. Five keys aren't so intimidating. The others are there, but dimmed out. You can still play all kinds of melodies. You can explore rhythm. You could play chords, though many won't realize it. Then you could switch over to a hexatonic blues scale mode: also self-contained; also not too intimidating. Then you could start displaying the notes the student plays on a staff above the keyboard. And so on. The skills learned on these limited keyboards are not abandoned when moving to more complex ones—they're fruitfully employed!
Each microworld will naturally engage different sorts of students, and that's okay: they should be free to roam between any within reach. The goal is to provide a sandbox—not a syllabus—for experimentation and the formation of understanding. Understanding is marvelous because it's so readily a feedback loop: we can use it to make more of it.
Chords sound rotten on an out-of-tune piano. I've written about why we think that might be: harmonics of the detuned lower notes end up a small distance from those of the upper notes, rather than directly overlapping them.
My piano was tuned this weekend. Playing it after the technician had left, the sound's transformation startled me: how rich and powerful my piano had suddenly become! I remembered, then, a less obvious but absolutely critical effect of correct tuning: sympathetic vibrations of a note's upper harmonics.
When you play a middle C, the resulting wave carries pitches of higher frequencies according to the harmonic series: C5, G5, C6, E6, and so on. The wave, emanating from middle C's string, will then excite the strings corresponding to the other pitches—providing the dampers on those strings are lifted (by pressing its key or the damper pedal). The sympathetic vibrations in these higher strings increase the volume of their respective harmonics, coloring the resulting tone.
In other words, not every C major chord composed of a C4, E4, and G4 will sound the same. The resulting sound will depend on which dampers are lifted in the octaves above! If the piano is out in tune, though, the frequencies of the strings in the harmonic series won't align, so there will be little sympathetic vibration.
Here's a little video demonstrating this principle in action:
Juggling exploration and exploitation in intuition
I spend my life building abstractions and automating processes. It's (literally) my job, yes; but my head's always whirring away to that end.
Say I'm tackling some piano piece. I transform the score's markings into musical language. I reify that musical language into instructions: "play this key here," "increase in volume," etc. I execute those instructions via complex musculature. Reacting to the sound produced, I form an interpretation which (hopefully) maps emotion onto the score via a series of fine adjustments.
I've spent many years automating these processes, relegating always more to my subconscious. I find even now that my playing of a passage will never match the sound in my head until it's memorized, and my hands navigate the keys unbidden. I build these automations by finding patterns and constructing abstractions. I might not be able to read a trio of notes at tempo if I address them individually, but I could recognize their collective shape on the page as an arpeggiated triad or an ascending scale, which I could then execute without further thought.
I suspect intuition is the automation of acts once executed consciously. Of course I've built most of this intuition unconsciously, but as I've begun to study studying, I've tried to exert conscious control of that growth. Cognizance has been the primary mechanism: constant reflection on and criticism of my actions informed by their outcomes.
Now I wonder: if I automate ever more layers of activity so I can focus my attention on higher levels still, can I ever stop scrutinizing the automated abstractions? What if a powerful new skill won't emerge until I change the way I've always intuitively performed some basic task? Maybe the skill tree's actually a maze, and I need to backtrack!
I've found new layers of understanding in my work1 that required me to abandon fundamental concepts I'd automated long ago. And in my piano education, I've spent years unlearning bad habits I developed as a child. So how should we balance exploration and exploitation here? It seems we can trust nothing, yet if we spend all our time questioning assumptions, we'll never draw any conclusions.
Including (software engineering alert!): fun with dependent, existential, and generalized abstract data types; invisible (to me) dangers in traditional asynchronous models; the ubiquitous and unformalized transactions and state machines in imperative logic; the value of immutable memory; knitting together the declarative, functional, and imperative; etc. Someday I'll write technically about software engineering again…
In the first chapter of The Beginning of Infinity, David Deutsch writes:
The role of experiment and observation is to choose between existing theories, not to be the source of new ones.
I have the most fun attacking questions with no satisfying theories in sight. So: how can we generate novel theories for a problem? It's critical to recognize that the creation of a hypothesis demands a wholly distinct approach from its evaluation.
Sometimes I uncover a critical idea by consciously exploring possible avenues, but more often, the realization flashes unbidden, with my attention elsewhere—like a lightbulb out of a classic cartoon. In these cases, I suspect my mind has been churning ceaselessly at some tree of possibilities beneath the surface of consciousness, projecting potential futures, searching my memory for relevant assocations, and evaluating the relevance of its findings before bubbling any useful notions up to my awareness.
We can propel this part of our psyche along by seeding and predisposing it for this pursuit. By providing kindling. I like to imagine waking up little subconscious drone ants and sending them in every direction, hoping that one will return with something useful later.
I find that doodles, in particular, provide especially effective kindling—whatever that means in your field: instinctive improvisation at the piano, whiteboard sketches of systems in engineering, lists of assertions in writing.
This process sounds much like brainstorming or mind mapping, but I don't think it is, exactly: those are aids to conscious invention and organization. I'm suggesting something more like rubber duck debugging, a technique in which the simple act of explaining a problem in great detail induces cognitive dissonance during the weakest parts of the explanation.
As a highly creative act, theory generation benefits from many of the same techniques that bolster artistic pursuits. Besides doodling, I've developed theories by exaggerating, by celebrating contrast, and by applying techniques from one field to another unrelated one. Whatever I can do to throw more wood on the fire.
In advice given to prospective students, my former piano teacher, James Boyk, suggests:
I'm not the teacher you want if you want to be passively turned into a pianist and musician instead of actively turning yourself into one.
This statement so concisely and wonderfully conveys my position on education and self-improvement. Its attitude absolutely informs the celebration of cognizance in my writing here, but so far, my suggestions in that regard have been for personal action. There's a role for others in self-awareness, too: my teacher might expect his students to take ownership of their education, but that doesn't mean he mentors passively!
The "1 on 1" meetings between employee and manager might strike you as a venue for second-person self-awareness, but these house tremendous opportunity for passivity: generally, the manager suggests how his employee might improve his performance. If only one could achieve mastery simply by following others' instructions! All we'd have to do is carefully note what our bosses say in those meetings and then execute those directions.
No, I suspect level-jumping changes originate from within. The teacher's role, then, is to inspire, to provide objectivity, to develop the student's taste and understanding so he can find his own weaknesses, and to make sure he can find the tools to defeat those weaknesses once he understands them.
Many advocate personal professional journals: every day, write down what you've done, what you're going to do tomorrow, what you could do better. I keep one, and I think they're a good idea. But, to crib again from Jim, a second-person journal can provide incredible perspective and self-understanding. After every lesson, his students email him a detailed summary of what they've done, what they're going to do, and reflections on the lesson's contents. He then responds actively to the student's self-analysis, calling out particularly effective or misguided ideas, and guiding their development. He's published some examples about halfway down this page.
We exchanged tens of thousands of words in this manner over two years, and I progressed all the more for it. And critically: this practice helped me learn to teach myself—the ultimate goal, since I would be eventually moving away. Now, I just wonder how I can apply this technique to my present pursuits.
This past weekend I encountered the following sales pitch for an industrial food service product: "are you ready to execute your soup program with excellence?"
Corporatese hilarity set momentarily aside, I derived inexplicable comfort from the vision of a mythical fast food franchise owner—ardent and passionate about every aspect of his business—who, reeling with the possibilities of this product, would engage and debate the salesman with wide eyes. Who might share his discovery with his friends (naturally: a set comprised mostly of fellow franchise owners), zealously contending the relative merits of competing industrial soup products.
I don't know if this owner exists, but even entry-level employees in my industry (software engineering) often behave like him. I suspect that's due to—for better or for worse—the exaltation of purpose in one's occupation: aspiring to do something objectively important, meaningful, or good during business hours.
I remind myself often how absurdly, outlandishly fortunate I am to have chosen an occupation with a supply–demand curve that makes purpose a conceivable job amenity. If I'd chosen another path, I'd be forced to spend those eight hours a day with utter personal indifference to my success or failure.
This celebration of purpose, though, comes with cost: if I think my work Matters with a capital 'M,' I've tied my self-esteem, temperament, psychic wellness to that undertaking. My day-time struggle with a nasty problem will seamlessly transition into detached night-time conversation, my frustrated, churning mind still back at my gray plastic desk under fluorescent light.
I now suspect the real danger is in a false sense of purpose: in accepting its bleed into after-hours mentally and physically, yoking it to consummate responsibility, equating personal success with the project's success—all while the work itself does not tally the imagined objective import, meaning, or good.
Sometimes this behavior is easy to spot. I worry about new college graduates choosing a first apartment a block away from work for a convenient commute. I'm self-aware enough to sharply change course when all my conversations somehow return to software. But unstated attitudes and priorities are harder to catch.
Having worked twenty-seven of the last thirty days, I've found I needed this brisk reminder: to attribute only as much meaning and intrapersonal attention to my work as to balance the purpose it delivers.
After learning how to season food correctly, homemade stock is the single best thing anyone can do for his cooking. A stock is a liquid extraction of complementary flavors, a broad palette upon which sharper splashes of color may be painted.
Stock is about balance and intensity of flavor. Making a mushroom sauce? If you were to just purée some cooked mushrooms, the isolated flavor would taste harsh, but add some stock and it will gain balancing sweet and vegetal notes. Those carrots need more flavor? Simmer them in stock instead of water: stocks are distillations of elemental flavors. They can be diluted or concentrated at will.
I credit stock with much of my ability to impress in the kitchen. They're as much cooking technique as they are ingredient. But, you keen, they're too hard to make! They take too long! Nay. Sit down. Let me tell you about my favorite stock. I make it almost weekly, and it's so easy that I am positively mystified why every cook in every household isn't making this elixir regularly. Let me tell you about vegetable stock, and let me tell you how to make it.
I made nasturtium soup this weekend, and the subsequent dinner's conversation remained imperturbably anchored to how unbelievably tasty it was. Yet it consisted of nothing more than nasturtium leaves puréed with vegetable stock, with a nasturtium flower for garnish.
Consider how else you might make this soup. You've got some nasturtium leaves: fine. How will you turn that into a soup? You'll need to add some liquid, clearly. Looking around the internet, I see a minefield of misguided ideas. Steep the leaves? That will diminish their brightness. No, they need to be puréed uncooked. But with what? Cream? That will dull the leaves' radish-y spice. Chicken broth or stock? The flavor of nasturtium is too delicate: you'll overpower it. Thicken with potato? We don't need filler here. Water? The result will taste thin and incomplete, lacking sweetness and umami.
Purée the leaves with vegetable stock and it will taste like a garden. It will taste like it smelled when you went back there to pick the nasturtium. It will taste like sitting in a park in spring. And then the horseradish spice of the leaves will give you a hearty, well-rounded slap and bring you back to the kitchen.
Making vegetable stock
Get three large carrots, an onion, two leeks, and a bulb of fennel. Peel the carrots and onion; discard all but the white and light-green part of the leeks. Wash the leeks and fennel. Chop everything up finely and put it in a big pot with a few glugs of oil.
Now you're going to sweeten up these vegetables by cooking them over low heat. You don't want to brown them—that will make a darker, richer variant of this stock, which should be bright and clean. Add a large three-fingered pinch of salt to draw out the vegetables' juices and turn the heat to low. Cook, stirring occasionally, for about fifteen minutes. Taste an onion. It should be quite sweet.
Add enough water to cover the vegetables, then add that amount of water twice more. Bring the water to a bare simmer and then let it cook for forty-five minutes.
Strain the stock. That's it. Store it in your fridge, but the stock will lose its brightness after a few days. That's why, when America's Test Kitchen set out to find the best grocery store vegetable stock, they concluded that all of them were unusable.
This stock provides a mostly neutral foundation for flavor. You can adjust for its intended purpose readily and improvisationally. Want grassier flavors? Add half a bunch of parsley. Heartier? Brown the vegetables while sautéing them. Asian flavors? Add ginger, a little lemongrass, and star anise. And so on.
How else you might use this stock
Reduce the stock by half, add salt and some fresh lemon juice; and you will have the best vegetable broth you've ever tasted. Add a bunch of fresh, perfectly-cooked summer vegetables immediately before serving and you will wow people. One friend, tasting in disbelief, asked how much honey I added to make it so sweet. I told him "three carrots."
Make risotto with it. Cook pasta in it. Any vegetable which you would normally cook by simmering in water (carrots, potatoes, turnips, pearl onions…)? Cook them in this with a healthy dose of salt. Make a fantastic sauce for any dish by puréeing a complementary, perfectly-cooked vegetable with reduced stock. Season it and use it in place of store-bought chicken broth: that stuff tastes like industrial chemicals.
Once you've got this down, no one will have to persuade you to learn about chicken, veal, pork, fish, and mushroom stocks.
Why does a perfect fifth sound so satisfying when a brash tritone is just a half step away?
It seems we have a physiological correlate to these psychological effects. As sounds arrive, the basilar membrane vibrates like an oblong drum head, stimulating tiny hair cells attached to nerves. Each hair cell corresponds to a perceived frequency, working like little band-pass filters.
One hair cell will be stimulated over some small range of input frequencies: for instance, hairs for middle C (262 Hz), will also react to sounds between 230 Hz (roughly A♯3) and 290 Hz (roughly D4). We'd say, then, that the critical bandwidth at middle C is around 60 Hz. Researchers have found that the bandwidth of a given hair equates to around a whole step on the western musical scale, basically the whole way up and down the keyboard.
As for the tritone: when a sound vibrates two hairs with overlapping critical bandwidths (like C5 at 523 Hz and C♯5 at 554 Hz), we perceive it as rough and rattling. Why are we wired that way? No study has conclusively determined the answer, but I'd wildly speculate that such sounds share acoustic properties with noise from friction. Rustling leaves; shuffling paws? Or screaming? Maybe it's just a warning sign when our brain can't readily separate two tones.
Whatever it is, this effect alone can't explain tritones' impudence. Take C4 (262 Hz) and F♯4 (370 Hz): they're several bandwidths apart, but they still sound dissonant when played together on my piano.
That's because when you play a C4 on the piano, you are also playing a bunch of other tones: 523 Hz (C5), 786 Hz (G5), 1048 Hz (C6), and so on for the next three integer multiples of C4's frequency. The tone, or timbre of an instrument—what separates the sound of a flute from that of an oboe—largely derives from the relative strengths of these extra tones (harmonics) to the pitch you meant to play (the fundamental).
So, your C4 (262 Hz) includes an inadvertent and somewhat quieter G5 (786 Hz). Note two of the tritone—the F♯3 (370 Hz) will include an extra bit of F♯4 (740 Hz). These two harmonics have colliding tiny-hair-bandwidths, just like the minor second we discussed earlier: hence the rough sound.
Even fairly consonant chords, like a major third, feature these collisions. C4 (262 Hz) and E4 (330 Hz) will clash at the third harmonic of the former (C5 at 1048 Hz) and the second of the latter (B4 at 990 Hz). Those harmonics are much quieter than the fundamental and first harmonic, though, so the major third grates less than the tritone.
Remember, though: the harmonics for a given note are produced at different relative intensities from instrument to instrument. This means that the same interval, played on different kinds of horn, could sound substantially more or less dissonant! Moreover, musicians can control their instruments' tone (i.e., in part, their harmonics' amplitudes) at will while they play. Sing an "ah", then raise and lower your soft palate. You'll hear your voice's tone change substantially. (And you'll sound ridiculous.)
Did composers subconsciously account for instruments' tonal effect on consonance when arranging their pieces? Is it possible that excellent musicians intuitively adjust the tone of their instruments to modulate the dissonance produced by the harmonies they form? Could this account for the sudden beauty of a chord in one ensemble's rendition of a piece you've heard a thousand times?
