Towards an Identity Model of Habits: Part I

My buddy James is a vegetarian.
I am not a morning person.
I’m a reader of fantasy books.

Remember those statements, ’cause I’m going to reference ’em later.

In the last few years I’ve been experimenting with various models of self improvement. Before I officially started this project I assumed that motivation was a significant catalyst for self change. After seeing it as a perennial problem (I can get psyched up for gym going starting on New Years, but it peters out pretty quickly, and the cycle repeats next year) I switched to other things.

I dabbled in gamification, because I saw its addictive properties as lowering willpower thresholds. Like motivation, it worked, but only for a while.

I’ve since focused on habits for the last two years, and though I’ve had a great deal of success, they’re only foolproof in relatively basic and linear behaviors. When things get complicated that paradigm just isn’t enough.

How are they not enough?
The linear model – what BJ Fogg advocates, of starting a Tiny Habit, reaching that hook point of automaticity, then naturally increasing difficulty, repetitions, or length of time until you achieve mastery – doesn’t seem to fully work all the time. Or rather it really falls a part when you’re pushing habits to mastery, which I see as another vector of effort (regimention/willpower and endurance/forming a habit being the other two vectors).

That vector involves plateaus in skill and the maddening frustration of constantly doing a task that is at least slightly above your current level.

It also runs into trouble when you’re dealing with families of skills. I advocate this not only because families can support each other, but in a world where time is of essence (we die, our bodies wear out), skills that have an accrual across time are necessary to start now to gain the benefits of daily minimums across time. If I start a habit of cardio 30 minutes a day, I may not master it. I might not get my goal of a six pack until I nail my eating habit. But for as long as I’m exercising, I’m accruing secondary cardio “points”.

Pushing skills in the vector of skill advancement throws a huge wrench into the equation because of habit harmonics. A dissonance starts – extra effort in one skill affects the solidity of other habits.

But the biggest problem with my current model is that it doesn’t attenuate in more complicated behaviors.

Let’s go back to the original three statements.

My buddy James is a vegetarian. When we go out and eat he avoids meat. In all scenarios. After the bars while tipsy and ordering pizza late at night, when going to a restaurant with friends with crappy vegetarian options, even in one place that had amazing pork tacos.

I do the same thing with fantasy books. It’s not as though I decide to read them – I HAVE to read them. It’s not even a choice. I need to have those few minutes before bed to scratch that itch and if I don’t have at least an option loaded on my Kindle, I start to get all itchy. The world is not right.

The inverse is important to analyze – I’m not a morning person. My waking up early is either a fluke or a deliberate preparation if I need it. Morning people are morning people because they enjoy it or they just are that way – it’s totally independent from fluctuating conditions. If they’re out late the night before, they still wake up early.

For all three – it’s an identity that’s welded in. It’s not what you do, it’s part of who you are, which not only makes it stronger, it also is able to somehow adapt incredibly well to changing conditions. Choice is also almost entirely scrubbed out of the equation.

For me this becomes an issue with eating and getting up early. All the other habits I consider foundational are easy. Working out – no problem, barring travel, it’s once a day at a certain time. Same with writing, meditating, and if I add flossing or recording finances. It’s a matter of if-then protocols – implementation intentions.

For eating that gets insanely complicated – it’s multiple times a day, across changing circumstances, etc. I believe it’s the reason I’ve had to scrap the habit several times, even when I’ve maintained it for close to a year. It just never stuck. And this is a big problem – eating is incredibly important for health, energy, and weight loss. It also has the biggest impact for whether I can socialize well later in the program – I don’t want to go out to meet people and, because of lack of willpower, blow out a previous habit of making good food decisions.

mask by 派脆客 Lee, tack by Zaheer Mohiuddin, welder by Per Hortlund

Recharging Habits

Since finishing NaNoWriMo my writing habit has been lax - automaticity has been difficult to achieve. A part of that has to do with improper transitional planning - I knew exactly what I had to do during my 30 day challenge, and after it ended I was left swinging in the wind. 

Another aspect is a lack of a proper implementation intention. It had been eroded by my recent travels, and for NaNoWriMo it had changed. My unstated implementation was “write A LOT” - which I did. But shifting gears to a normal schedule my if-then protocol was gone. And this is very noticeable when compared to my very new rowing habit, which has a crisp if-then (as soon as I get up, I row), a fact that’s reflected in rapidly soaring SRHI scores.

I switched up my routine yesterday and today, immediately writing after rowing. It just feels more automatic. It appears that the closer I have a task to waking up, the more charged the habit gets. Why? I think it has to proper implementation - the further a long in the day the more willpower stores are depleted. Also more tasks come up later in the day. I need to eat, I need to go to the bathroom, I need to cook. All of those tasks are not precisely pinned down - they change, making the implementation sloppier.

That usually doesn’t matter so much - but after various forms of degradation (travel, a 30 day challenge, getting sick), it starts to make a big difference in automaticity. 

A while back I talked about the potential that all long-term habits may need a “re-charge” once and a while. Scott Young, in his post “Why is it So Hard to Create Permanent Habits?” describes this train of thought.

In the post Young talks about how many habits have to be restarted. We want to think they will be permanent, but they often aren’t - habits for him are a medium-term strategy. They are, in his terminology - “metastable” - they lower thresholds of action in some ways, but not all ways. And because of this, they often have to be restarted depending on the changing action you are doing in the habit. 

This idea of metastability conforms to my experience as the reason why I’ve found few habits have had permanent lifespans. Inevitably, the habit breaks down because of a temporary lifestyle change: a vacation, an illness, needing to move or work overtime. These create shocks which are often enough to break the behavior, increase the decision cost, making it no longer automatic when you return to the habit.

His full post is really interesting, and I’d like to analyze it fully in a separate post. I agree that shocks will destabilize habits. But I think proper mid-range planning can compensate allowing you to “shelf” some habits at lower daily minimums (which he mentions) or, in this case, “recharge” them by rotating them in a daily regiment. 

I also think that tempering a habit comes with these periods of unstableness - there’s my quarter mark theory, there’s a dip in the graph before a superhabit is formed - without a metric to determine habit strength or a habit of tracking habits it’s hard to see whether a habit is lost, or if it’s still there and going through a weak patch.

How To Form an Eating Habit Revisited

A find it interesting that though there are many articles on how to form a habit, there seems to be few articles on how to take the scientific approach of willpower and habit formation and apply it to eating.

In my NaNoWriMo book on this project, I took some time and theorized on how I’d be able to do this if I were advising someone starting from scratch.

To do this there are a few protocols I’d apply:
1) Implementation Intention
2) Sequenced habit chains, or “bookending”
3) Start small - “TinyHabits”

Based on this, my advice would be to  start small - and one small start is focusing on one clean meal a day. I’d also advice to make this automatic by having a clear cut implementation intention that’s in a chain of habits. I start out my days rowing, then I take a shower, then I meditate. I’d tack this on to the end of that chain.

The problem with my habit as it stands is that it’s fuzzy. My habit is essentially “eat clean.” That doesn’t really promote automaticity. Automaticity grows from having a clear time or sequence - an “if-then” parameter in the  mind. By not having it clear cut it promotes confusion - an unclarity in the forming habit.

It’s also all or nothing - if a Tiny Habit is ludicrously easy so that you feel like you’re more likely to do it, then having an all or nothing approach doesn’t really promote clean eating. It can’t be Tiny.

Nor does it incorporate how I prep for eating.  A nested habit would be beneficial for this- something like - on Saturday I go shopping for the week, and then I eat. 

My advice would be to fully master this as a habit, then move on to the next step.

The extension would be really difficult because there’s not really a bookmarking point for, say, eating at 6 pm. It would have to be an if-then based on time of day.

If I were talking to someone else, I’d of course start with smaller steps - removing soda, for example. But that’s not really something I have a problem with anymore - I generally drink almost nothing but water. 

I was talking to Lydia about this and she presented a counter argument. Articles have come out that suggest that things like gluten and sugar act almost like heroin in the brain, causing us to want to eat more. Her question was - would you apply this strategy to a heroin addict, or would your first step be to have them replace the drug fully? In methadone clinics heroin is replaced, then cut down.

So the analogy would presumably result in replacing wheat/sugar with substitutes, and then lowering the substitutes. Of course this is only from the Primal point of view.

I really don’t have a solution to this, except by looking at the past, seeing what I’ve done, and seeing how I’ve failed. I’ve gone the all or nothing approach, and it has clearly failed. My tendency now is to do something different, which seems to be to try to the piecemeal approach.

I do know that automaticity for my clean eating has been all over the place, and at least a portion of it has to be because I haven’t used the tools for habit formation at all in this particular habit. And it hasn’t just resulted in “fuzziness” in eating - it actually has a tendency to mess up other habits.

When I’m not pro-active about eating (pro-active being striking at a prearranged time versus “whenever I’m hungry) eating becomes an interruption in the habit chain of my morning. I believe that striking first allows me to incorporate it into the fold. And if I do this, I have yet another solid portion of the chain to implement another habit - writing or recording, both of which have been adversely affected.

I definitely think that other techniques I’ve used have really helped - especially the Flash Diet, which supercharged my eating during my 30 No Bread Challenge. I feel this can be incorporated into my progression.

The Weekend Power of Habit

I endeavored to take this weekend off. 

I automatically did all my habits on Saturday. And I did several of my habits yesterday (Sunday). There was a point where I was out with a friend and we had planned to go to a place that had only clean meals. We instead went to a Tex Mex place (my kryptonite). 

I automatically ordered a salad. It came completely covered with fried tortilla strips. And I just automatically pushed them to the side and ate everything else.

This is actually working. It’s something that I don’t really believe because the results aren’t immediate. But it’s in moments like that where I begin to understand that my self control is shooting through the roof - that’s definitely no the person I was 1 year ago.

My weekend in general was incredibly relaxing. I am starting to really focus on regimentation and am now really convinced that the art of regimentation is a significant and important part of this project. I think it’s something that most people have problems with - the ability to be completely caught up in something and just, in the next moment, completely forget about it. I think an inability to do this results in massive leaks in self discipline across time.

A New Protocol for Work, Pavlov/Click Training Style!

A New Protocol for WorkWork just isn’t progressing well enough, and I think it’s mostly because I don’t have a solid trigger.

Now normally my triggers are sequential ones, using bookending. I have my stable daily event - waking up. I wake up, I take a shower, I meditate, I do my bodyweight training, I write, and….by that time I’m a little mentally tired.

I end up dithering online for a while. Making coffee. Reading emails, whatnot. I need a bit of a break.

Earlier I had thought about incorporating a unique sound, just like Pavlov’s dog, to become a trigger. I rejected a time based alarm because my schedule changes day to day. But what if I did a timer and set a specific sound?

So today I started this. After I finished writing I gave myself 25 minutes

I selected a unique sound - “Waves” - to set my habit to.

After 25 minutes (the pic shows it counting down) my timer went off and I immediately started opening all the pages I need open for work. Now it’s “crisp” - I’m tying it to this particular unique sound at my choosing.

I think this will work, but we’ll see. My minimum is still set at 20 minutes of work. I finish, record and write on this blog, then finish up with the rest of work. But what if I set each time I start on work to the same ring tone? The trigger is still there, but now it’s going off twice, maybe multiple times a day.

My hope and hypothesis is that I will be able to really vanquish this habit in record time because of this. Why? Because I’m practicing the habit multiple times a day, it becomes triggered more often, and automaticity gets practiced more. Frequency shoots through the roof. Length of time is already maxed out, so the only thing that would need more points is identity questions.

We shall see!

A New Plan for Habit Progressions Part 3 - A New Hope

What I’ve done for the last 38 days is work solely on 750words.

My theory is that after the a rough patch and after the midpoint, things start their inevitable slide towards automaticity - this is a theory that runs counter to Lally’s research, and is something I’ll discuss later.

My new plan was to work solely on 750 words until after the midpoint, and then start a new habit, the theory being that after the midpoint the will needed to sustain a task becomes less and less. Two tasks overlap, but in the best way possible. The starting task is easy for a bit, and the previous task is approaching automaticity and taking less willpower.

I feel the change in 750 words now. I’ve streamlined the process using the book Self-Discipline in 10 Days and pegged it to a specific time i.e. as soon as I get up in the morning. I’ve taken the Automaticity Index questionnaire and it seems to corroborate my feelings - it is on it’s way to becoming a habit.

I chose working out as my next habit - specifically doing some form of kettlebell workout right after I finish my 750 words - and I started Day 1 this morning. This will be a long one - I’ve pegged it at a 125 day midpoint with an estimated 250 days before it becomes a habit. I’ll be streamlining the process like 750 words, and I’ll be taking notes on how much willpower it takes, it’s automaticity by taking the questionnaire, and how it interacts with my previous habit in the making. I’ll also be using fitocracy and a Google docs spreadsheet to keep track of it.

A New Plan for Habit Progressions Part 2 - Assessing Tasks

So I’m going to just attempt a rough estimate to peg ½ points of automaticity based on difficulty. I’ll divide my tasks into 3 levels
Easy - Reading (Goodreads), Daily meditation, flossing
Medium - Khan Academy course, Duolingo, 750 words
Difficult - Working out, eating right, not drinking alcohol or coffee, music

Flossing is a simple task, reading is something I like and do regularly anyways, and daily meditation is itself very simple and takes very little time.

Courses tend to take more energy from me - especially if they require homework, and 750 words though is enjoyable, sometimes takes a long time depending on the day

Working out is difficult even according to Lally’s data, as is eating right over time. Alcohol and coffee are such an easy thing to indulge in here and there because it tends to be cemented with socializing.

Two things I didn’t include are social events and quantifying my finances. Socializing is hard to start off, and I think it should be left until the end, especially since it’s often difficult to eat right - so it should be done after automaticity is achieved with that. And Finances….I have no idea how difficult that will be especially since initially it would be just a matter of observing my finances, not doing anything about them. So roughly speaking, I’ll rate the actions from 1 (easiest) to 10, then roughly rate the number of days to reach automaticity, and then half it.

Flossing - 1 - 18 days - 9 day midpoint
Reading - 2 - 30 days - 15 day midpoint
Meditation - 3 - 40 days - 20 day midpoint
Finances - 3 - 40 days - 20 day midpoint
750 - 4 - 50 - 25 day midpoint
Duolingo - 5 - 66 - 33 day midpoint
Khan Academy -  5 - 66 - 33 day midpoint
Music - 7 - 100 - 50 day midpoint
Not drinking coffee/alcohol - 10 - 250 - 125 day midpoint
Working out - 10 - 250 - 125 day midpoint
Eating Right - 10 - 250 - 125 day midpoint
Socializing - ?

I’ll have to take into account that these numbers may change as I evaluate them with the automaticity index.

A New Plan for Habit Progressions Part 1

So in light of feeling the brunt of all these attempts at habit formation, I’ve stepped back and tried to look at a better plan based on all the stuff I’ve been reading.

So here’s my theoretical framework: If self discipline is one deplete-able resource, and if automaticity is when an action approaches an asymptote of 0 willpower, then the best way is to have a slowly unfolding progression of new actions taking over as the first action slides towards that asymptote.

For example, if drinking a glass of water before breakfast takes 18 days to form into a habit, then I’d want to start another habit at some point past the mid point where the slide to full automaticity is underway.

I’m assuming, perhaps erroneously, that tasks take a certain amount of willpower on day 1, and start getting harder as the “streak” of sustaining the habit continues as days go by. This certainly is true for my own personal attempts. Going to the gym is easy on day 1 or 2, but it gets more and more difficult to continue the streak of days (obviously there are some random days where it’s more or less difficult).

Given the above, and given that different tasks take different amounts of times to reach automaticity, then it makes sense to start Habit 1 until the midpoint of automaticity, then start Habit 2 until it reaches ITS UNIQUE midpoint, etc. This way the process of automaticity itself shoulders the brunt of sustained will - rather than how I’ve been doing it now - which is simply adding tasks 1 week later regardless of their difficulty or willpower expenditure.

So now the question is, how do you establish midpoints? I do not want to test each task, so what I’ll do is ball park it using Lally’s experiment - I’m assuming that tasks that take the least willpower will take the quickest to form into habits and the ones that take more will take longer.

So with Lally’s data, I can estimate that easier tasks will take 18 days, middle of the range will take 66 days, and the most difficult will take 250 days - I’ll divide each to peg the midpoint, and estimate the variations between tasks that I’ve already labeled as Easy, Medium, and Difficult.

Thoughts on Interludes and Automaticity


“Interludes” is what I’m calling those points where your interrupted - the inevitable trip, friend coming into town, or event that disrupts your careful practices.

My mom and I were talking about this whole project, and one thing that I’ll have to confront is how to deal with these moments. Obvious one solution is to keep carrying on. But in some cases this is genuinely not possible due to time restraints.

I’m curious if another solution would be a partial but time constrained execution of a skill.

For example, if your goal is to write 750 words a day, would writing 20 words work? For working out, would doing a few air squats that take 5 minutes work as a placeholder for the full execution of the task?

According to Lally’s experiment, missing a day or two here and there didn’t seem to adversely affect the approach of aromaticity. But would a half execution be better?

Notes on Experimentation

Lally’s experiments on self-discipline got me thinking about experiments I could possibly perform on the subject. Here are a few thoughts:

1. If there is an automaticity index, is there a happiness or willpower index? If not, is it possible to com up with them?

2. Does mood affect willpower? It seems like it does given my experience with the effect of humorous videos on ego-depletion.

3. According to Lally’s study, different actions take different amounts of time to develop into habits - Do actions that take less willpower to develop automaticity take an equal amount of time to lose automaticity? Or Do they take more time?

4. What other actions rejuvenate willpower?

5. Does relaxation affect willpower?

6. What is an optimal willpower workout?

7. How many actions are optimal for habit formation? Given 6 actions that have a similar “will power” and automaticity value, how many actions are “stackable” - what is the likelyhood that 2 vs 6 actions will achieve full automaticity.

8. Is there an automaticity constant that can be derived from specific actions? If automaticity can be mapped, and if will can be mapped, then shouldn’t there be a value derived for willpower and automaticity for a given action?

9. Do actions that rejuvenate ego-depletion have the same affect or do they diminish? Can I exert willpower, then rejuvenate willpower, and repeat ad infinitum? Or does each rejuvenation yield fewer and fewer “willpower points?

These are all rather vague notions, but I think it’s important to note them so that they can be examined in detail later.

Phillippa Lally and the Number of Days to Form a Habit

Common consensus seems to suggest that it either takes 21 days or 30 days to form a habit.

Based on my own observations, this is clearly not true - and Philippa Lally’s study in the European Journal of Social Psychology seems to suggest otherwise.

The study analyzes a group, has them perform various tasks, and asks them at what level of of automaticity they are at with these tasks every day.

The study suggests that some tasks take less time and some tasks more time before they are considered a habit, but the average is 66 days.

This brings up a number of interesting points:

1. Perhaps I should use an index like she did in order to assess my level of habituation for any given task.

2. Perhaps programs like Level Me Up, which use a number of hours to determine mastery, are not as good as a program that functions by checking out a daily practice of a skill. I’ve already begun to have doubts about Level Me Up - this might be it’s death knell.

3. What are the tasks that are easier and take longer to form into habits? I’m assuming this has to do with those that take more willpower. If habituation can be indexed by a questionnaire, perhaps I could come up with a willpower index in order to attempt to find some sort of correlation between willpower expenditure and “automaticity” - what the studies are using as the technical term for the progression of habit formation.

This is some really exciting stuff, because it’s the first real testing as to the mechanics of habit formation in scientific conditions.