Reference Points
I just spent some time reading Thomas Schelling’s “Choice and Consequences” and I heartily recommend it. Here’s a Google books link to the chapter I was reading, “The Intimate Contest for Self Command.”
It’s fascinating, and if you like LessWrong, rationality, understanding things, decision theories, figuring people and the world out—well, then I think you’d like Schelling. Actually, you’ll probably be amazed with how much of his stuff you’re already familiar with—he really established a heck of a lot modern thinking on game theory.
Allow me to depart from Schelling a moment, and talk of Sam Snyder. He’s a very intelligent guy who has lots of intelligent thoughts. Here’s a link to his website—there’s massive amounts of data and references there, so I’d recommend you just skim his site if you go visit until you find something interesting. You’ll probably find something interesting pretty quickly.
I got a chance to have a conversation with him a while back, and we covered immense amounts of ground. He introduced me to a concept I’ve been thinking about nonstop since learning it from him—reference points.
Now, he explained it very eloquently, and I’m afraid I’m going to mangle and not do justice to his explanation. But to make a long story really short, your reference points affect your motivation a lot.
An example would help.
What does the average person think about he thinks of running? He thinks of huffing, puffing, being tired and sore, having a hard time getting going, looking fat in workout clothes and being embarrassed at being out of shape. A lot of people try running at some point in their life, and most people don’t keep doing it.
On the other hand, what does a regular runner think of? He thinks of the “runner’s high” and gliding across the pavement, enjoying a great run, and feeling like a million bucks afterwards.
Since that conversation, I’ve been trying to change my reference points. For instance, if I feel like I’d like some fried food, I try not to imagine/reference eating the salty greased food. Yes, eating french fries and a grilled chicken sandwich will be salty and fatty and delicious. It’s a superstimulus, we’re not really evolved to handle that stuff appropriately.
So when most people think of the McChicken Sandwich, large fry, large drink, they think about the grease and salt and sugar and how good it’ll taste.
I still like that stuff. In fact, since I quit a lot of vices, sometimes I crave even harder for the few I have left. But I was able to cut my junk food consumption way down by changing my reference point. When I start to have a desire for that sort of food, I think about how my stomach and energy levels are going to feel 90 minutes after eating it. That answer is—not too good. So I go out to a local restaurant and order plain chicken, rice, and vegetables, and I feel good later.
Schelling talks about in Choice and Consequences about how traditional economics applies a discount rate, but how that fails to explanation many situations. Schelling writes, “[The person who] furiously scratches would have to be someone whose time discount is 100% per hour or per minute, compounding to an annual rate too large for a calculator.”
Schelling raises more questions than answers. But I think one of the answers is clear, and that answer is reference points. The man who scratches his rash at the expense of a much worse condition immediately isn’t discarding the future. He simply isn’t referencing it when he makes his decision. He itches, he references scratching with an immediate abatement of the itch.
Eliezer writes in the theory of fun that to sell an idea to someone, you usually don’t need to convince them it’s a good thing to live with for their whole life. You only need to convince them that the first hour or day after they choose is going to be good.
And… I think that’s scary, because it’s true. People reference the immediate <em>very</em> short-term consequences of their actions, instead of the broader pictures. Whether that’s exercise, junk food, scratching a rash, buying a bigger TV, or conceptualizing eternity.
This explains a lot of why people act the way they do. It also explains a way forwards for you—gradually evolve your reference points so that thinking of junk food is thinking about feeling that heavy weighted-in feeling in your belly and so that exercising is the rush of good hormones and pleasantness of a good workout. Imagine scratching a rash as doubling your discomfort instead of abating it and imagine how incredibly nicer your future surroundings if you save and invest that money for just a short time longer.
Your reference points establish how you value things. Change them, and how you value things will change.
I just switched my reference point for reading LW right now from “enjoying interesting thoughts of smart people” to “staying up too late, feeling guilty and logy in the morning”. G’night folks.
I think you accidentally deleted something in the midst of editing, because this paragraph comes out of nowhere. Discount rate for what? Scratching? You hint at the context in the next paragraph but unless I missed it you don’t actually explain anywhere the example that you’re referencing.
Also, your didn’t work. :)
The actual content is interesting, though. It’s trying to remind me of something else I read recently but I can’t quite put my finger on it.
It seems to be missing something, yes. I understood it refers to scratching an itchy rash and thereby making the rash worse, and likely to itch worse in the future. Before the rash context later on, I thought it might refer to scratch card lottery as well.
The economic model is that your utility function (over world histories) is the integral over all time of a quantity that depends only on the instantaneous state of the world times a quantity that depends only on the time index. Typically this latter quantity is of the form e^(-kt), and k is called the discount rate.
I tend to be much more motivated and disciplined than most people, and seldom have problems with “akrasia,” especially in the past several months (quotation marks for not separating reality at its joints). I also tend to have unusual reference points, and I sometimes consciously work to change them. These unusual reference points seem to correlate with my best experiences of real motivation, and I think this technique has been very helpful for me. A few examples:
I associate exercise with its result of being more in shape and tasks becoming easier, taking less effort to lift things, being able to run further without being winded, etc.
Back when I was studying Japanese intensively, I associated it with in image of having mastery over the language and all its characters, and strolling about some neon-lit Japanese street with some wonderful Japanese woman.
Since childhood, my reference point for food has been unusually health oriented, and eating healthy is not difficult for me. I similarly found it relatively painless to become vegan, gaining a reference point of an absence of terribly treated animals, and feelings of enlightened benevolence.
Difficult but still worthwhile tasks are associated with ultimately gaining impressive abilities that most people don’t have because they haven’t put in the requisite efforts.
Interesting. Most of your examples fit a hypothesis I have, but there’s not enough information for me to confirm or deny it for this one:
What I’m wondering is, what specifically do you find pleasant about those results? (There is an emotion cluster that’s alluded to in all three of your other examples, so I’m curious if it applies to the fourth as well.)
I have something of a supergoal to become increasingly competent and capable, very much enjoying those things for their own sake (along with some instrumental benefits). I also have some terminal desire to be “cool”, which is very related but distinct. The appearance benefits mostly just serve that goal.
Was there an emotion cluster besides this one you were picking out? It sounds potentially enlightening to hear it.
Nope, that’s the one. I’d been hypothesizing recently that such aspirations would be the positive emotions most likely to lead to sustained and disciplined actions.
[I group emotions (and the evolutionary terminal values driving them) into four major categories: Significance, Affiliation, Safety, and Stimulation—SASS for short. The desires to be cool, impressive, admirable, and/or proud of one’s accomplishments all fall under “Significance”. While each group can motivate both positively and negatively in principle, each seems to impart different behavioral biases in practice. Significance and Affiliation seem to be the only ones that can motivate long-term disciplined behavior, and Affiliation seems to need to be other-directed in order to be useful.]
Interesting. That analysis makes a lot of sense to me, and I largely agree with it, including the idea that Affiliation and Significance are the most useful for long-term disciplined behavior. However as I was thinking back on this the other day, I realized that my second example:
was a drive for stimulation. None of my friends or family would have been particularly impressed (or unhappy) with such an accomplishment, and I myself didn’t consider it that amazing. Many people on Earth know 2 or more languages fluently, and while the people around me in Japan would find me a novelty or somewhat impressive (partly because of the cultural subconscious idea that Japanese is impossible to learn), I would basically just be “able to speak the language”; proficient but not exceptional. I also wasn’t thinking of doing anything really advanced like becoming an author in the language, just in getting by with it and understanding the language around me. My drive in this case came largely from this feeling of Japan as a sort of magical place, based partly on experiences I’d had during an earlier visit of 2-3 weeks. That really was a magical (and occasionally frustrating) trip.
I think this is just an unusual exception to the generally good analysis you put forward, but for accuracy I wanted to point it out.
Thanks!
I think you’ve hit upon something here. One of my many akrasia-enabling habits is a free game called League of Legends. Each game is a team game where players control a single champion (picked before the match starts). You acquire gold in the game by killion minions, monsters and other champions, and you use that gold to buy items which increase your champion’s stats. The game ends when one team destroys the other team’s base.
This game follows the DOTA format, and if you’ve ever played DOTA you know that the games can sometimes be quite engaging. Unlike the original DOTA, however, in League you can level up your account. You have to play a couple hundred games to get your account to the maximum level, where the full range of runes and mastery points become available to you. Also at level 30, ranked games become available to you, which is an entirely new tier of competitiveness. Each game is typically around 30 minutes, but they sometimes go an hour or longer.
With all this in mind, you might wonder why anyone with interests in anything other than gaming would take the time to play this game. The answer is that sometimes the games are genuinely riveting, back and forth battles of wits and strategy. There are over 60 different champions with many unique abilities each, so games are often quite different, and by combinatorics alone it’s difficult to exhaust all the possible games.
However, even that last paragraph of glowing description shows that my reference point is currently optimized for time-wasting. Genuinely fun games are rare because people sometimes leave the match, making it 4v5 in favor of your team or the other. This often results in a boring or frustrating game, because it’s no fun to crush another team 5v4 and it’s certainly no fun to lose 4v5. Even when people don’t leave, sometimes your teammates or opponents are just bad: their item build sucks, or they don’t know how to use their abilities effectively, etc. No one finds it fun to lose because of bad players on your team, and I personally don’t find it fun to mercilessly crush unskilled or inexperienced opponents.
By changing my reference point to think explicitly about how often League of Legends games are genuinely boring or frustrating and not thinking about the rare times the game is fun, I should be to curtail some of my game playing. I hope anyways.
Edit: A quote because it seems to be relevant to the notion of changing your reference point:
-- Alan Kay
+1 - please, please post that to the Rationality Quotes thread.
Getting the right reference point is what I use to make up my IQ deficiencies (a mere moderately bright 130ish as opposed to the typical LessWrong levels) so I know roughly what I’m talking about around people smarter than me. Fortunately, it is in fact one of my aptitudes, and people smarter than me seem to like hanging out with me. And I’m not even that pretty.
This reference point phenomenon is, to me, the kind of thing that seems obvious after you’ve already done it, but isn’t actually helpful if you’re trying to change a behavior.
If you’re trying to get into the habit of going to the gym or whatever, you already know that it’s going to be to your benefit in “far” mode but “near” mode you just doesn’t want to go. Near mode you has better stuff to do right now, healthfulness is far mode’s problem. You can’t re-program yourself to associate “working out” with “feeling good” until you’ve already been doing it for a while. This has been my experience, anyway. I run every day, and it’s just part of what I do, but the catalyst to getting into this habit wasn’t that I was suddenly able to convince myself that this was something that was good for me and that later on I’d enjoy it, even if I didn’t enjoy it now—no, the reason I started running was because at the time I had an immediate desire to do it (stress, pent-up frustration with life situations). I have absolutely no ability to trick my near mode to do things to the benefit of far mode; it has to have utility to me, right now.
Of course, now that I’ve been doing this for a while, when I’m about to go run I don’t even have a mental dialog where I have to convince myself that it’s something that I want to do—I just do it. If I haven’t run today, then obviously I am going to run, there’s modus ponens. If for some reason I have a voice saying I don’t want to do it, my brain immediately overrides that with, “But that just doesn’t make sense!”. If I were trying to convey this mental process to someone else, I might say something like, “well, I just envision myself running and having a good experience, and then not running and not having that good experience, so I’ve changed my reference point”. This after-the-fact explanation sort of explains what’s happening in my mind, but doesn’t actually give somebody else tools that allow them to actually copy it. The only advice I’d give is to find an actual compelling reason to do it whatever it is right now, rather than trying to fake yourself into thinking you want to do something that you really don’t.
Basically, you’re right about the changing reference points but I think you’ve got the order mixed up. That happens after you’ve changed the behavior.
Mm… I think I disagree, at least somewhat.
There’s a bunch of distinct clusters of mutually reinforcing mental states in my head, some of which increase my likelihood of doing X (e.g., run) and others of which decrease it. It seems reasonable, and is consistent with my own experience, that deliberately and vividly imagining the pleasant experiences associated with doing X activates the former and inhibits the latter.
That said, I think different techniques work well for different people in this area, just like people have different learning styles and visualize to different degrees and etc. If deliberate/vivid imagination isn’t something I’m good at, for example, this technique won’t work well for me, and I’ll do better with some other approach. So I’m sure you’re right for some people.
Tangentially: one of the most effective things I ever did to lose weight was experience PTSD-related hypervigilance around food and exercise. I’ve mostly gotten over it, and am far more likely to snack and slack off on exercise now. I wouldn’t trade back if I could, but I have to admit it worked.
FWIW, according to these guys, there isn’t actually any good evidence for that:
I agree with your point—that your technique has the potential to work better for some people than others—but I thought you might like to know this.
You know, the beauty of this is that as I wrote that line, I thought to myself “Huh, I wonder if that’s even true. On this site, I should probably look it up before I use it as an analogy. Naaah, never mind… I know enough educators who swear by it that it’s probably true enough.”
Clearly, I should be listening to my niggling doubts more.
Thanks for the correction, and the pointer.
Hahaha. Well, to be clear, we’re not sure it’s NOT true—and if assuming it has had good results for many educators, it seems more likely than not that there’s at least some grain of truth to it—it’s just that it hasn’t been tested properly yet. :)
My expectation is that any theory that encourages educators to pay attention to individual nonhierarchical differences among their students will have good results relative to any theory that encourages educators to treat all students either as identical or as stratified, simply because teaching is a social relationship among humans and social relationships tend to benefit from mutual respectful attention.
So I’m reluctant to choose one theory over another solely on that basis.
Hear, hear. The difficulty of that is not to be underestimated, though. I had a teacher once that I liked a lot—I found his explanations really clear and interesting, and he was a funny and nice and interesting person himself. Chatting with some of his other students outside the classroom, I found that everyone I talked to said something along the lines of, “I feel like he teaches in just the right way for me to understand.” So either he’d found the magic teaching style that works for everybody, or he was simply giving us the same information in several different ways throughout the class, and each of us only remembered the one that worked. It was pretty cool, and I have tons of respect for that guy.
(For what it’s worth, what I remember best about his classes was how visual and hands-on they were, despite being about relatively abstract topics. For example, when we were learning about ext2 filesystems, he put pieces of paper with file, directory, or inode contents on the various desks, and had us scurry around tracking down specific information in files—basically, learning how to be a shell! Or, similarly, when discussing standard i/o filehandles, he held up a physical box with one wire going in and two wires going out. It didn’t matter that the box itself wasn’t actually doing anything; seeing it made the metaphor crystal clear.)
In my experience it does seem to be true. In Aikido class, some people did seem to pick up things easier from seeing, others from hearing, etc. One thing I noticed was that my brain doesn’t really get ‘left’ and ‘right’ - ‘same side’ and ‘other side’ make much more sense. Don’t people who learn things in ‘left and right’ have to learn every technique twice, once for each side? (Strangely enough, it appears they don’t!)
But learning things better one way doesn’t mean other ways are useless. All ways reinforce each other, and now that I’m learning Alexander technique I notice that sometimes I get a ‘click’ from spoken directions, other times from touch, etc. If you’re a teacher, you should probably use as many modes as possible.
This reminds me of a story from a friend who works in a bike shop. A customer came in to buy some pedals, and asked of the two employees (my friend and her coworker) who were standing nearby, “Do I need to screw one of them in differently when I install them?”
Speaking at the same time, one employee said, “Yes, both of them screw in forwards” and the other said “Yes, the thread on the left pedal is reversed.” The customer’s eyes glazed over and one of them quickly picked a single way to explain in more detail. The point was that, given three accurate mnemonics for the same thing, you gotta just pick the one that makes sense to you.
I hadn’t heard of the Alexander technique before and just looked it up. That seems really interesting. How are you finding it?
Bonus points if they described the concept of chirality. Who knows, maybe the customer will need to learn organic chem some day!
Sorry for the late reply, but I only just noticed the little red envelope. :( I’m coming up on my last Alexander lesson in a few days, and I really like it. It’s not (yet) been as life-changing as I’d hoped (I still have bad knees (sigh), and people don’t suddenly react very differently to me), but it has improved my posture a lot, and I’m very happy with that. At the very least, it’ll help me not get RSI or back problems.
I only just noticed this reply, so we’re even. ;) Thanks.
But since I can’t actually copy this technique and have it work every time, I suspect that other people find it equally unenlightening, which is why I think it’s a poor model for actually bringing someone out of procrastination. That is, I think there’s something else going on in your head in addition to just imagining the pleasant experience that you’re not recognizing and therefore can’t communicate. Not just you, of course, this is exactly what I’m struggling with: identifying why my brain works differently some days than others. I’m in the middle of tracking what the conditions are when I have an “on” day versus an “off” one. I’ve already noticed that if I write down the patterns of thoughts that I have when “on”, thinking them back to myself when I’m “off” doesn’t actually change my mental state. I really want to identify what factor(s) will turn me from “off” to “on” every single time. An impossible goal, alas.
Do you think that on certain days, you’ve had to draw on your resources a lot in order to manage interactions with other people (say, pushy salespeople or whiny coworkers)? After a tiring day, or under conditions of stress (hunger or lack of sleep, for example), your brain would definitely be working differently, and not necessarily to your best advantage in the self-control department.
For me, the challenge is sometimes recognizing that I need to actually stop pushing myself.
This matches my own experience with exercise pretty closely. To be honest, I’ve never really learned to enjoy working out, but this hasn’t posed a particular barrier to getting in shape. It’s simply become something I do, and would feel uncomfortable not doing, like brushing my teeth. Once you cultivate habit, a new outlook may follow, but a change in outlook is not strictly necessary to conserve willpower.
I think that’s a rational response, by the person you’re trying to convince. The more distant the promised reward of whatever someone is touting to me, the more it will cost to reach it, and the more convinced I would have to be before taking it seriously enough to even begin. Or to put that the other way round, show me something that I can do right now and experience evidence that it works, and the bar for you to sell it to me is much lower.
“Jam tomorrow” is the promise of crooks and charlatans.
In the timespan under discussion
you just justified crack usage
That was, I believe, implicit in the original post. The reason that it’s scary is that it doesn’t just apply to scratching a rash, it also applies to doing hard drugs.
Reference points, as used by default, make it very easy to throw your life away. That’s what makes it scary.
“Jam now” is a necessary condition, not a sufficient one.
If I decide to, say, learn Spanish, then I expect every hour spent on the task, including the first one, to pay a perceptible return.
I accept this as a valid point—first hour/day is an important heuristic indicator of goodness, but Eli wrote
Or at least, that’s how it invariably worked in pre-legal societies. No wonder we’re mistrustful of delayed payoffs.
I often use childhood as my reference point. I wanted to be out of school, I wanted all the time in the world to play computer games (or to read science books), I wanted to learn more, and I wanted my parents to buy all my toys for me. And I know I can achieve all of them now. I just happen to have developed new desires, and I become unhappy when I can’t fulfill some of those new desires (like getting a significant other, or getting into a prestigious university). And then i ask myself, “well, really, shouldn’t I think back to how much better I am compared to myself when I was 11?” It’s not a realistic reference point for everyone, though, since desires change with puberty. But even then, if you use age 15 or 16 as a reference point, can’t you see that you have much of what you wanted back then? (although friends are often harder to get than before, for some people anyways) And that you probably now have so much more than you had back then?
I’m not convinced that this amounts to practical advice. As you eloquently argue, the whole problem of akrasia is being unduly influenced by immediate consequences. The way I would put it is that immediate consequences act like a drug that distorts your decision-making. So “change your reference point” sounds like “don’t be influenced by this drug”. Easier said than done!
There’s a parallel debate about this going on here: http://lesswrong.com/lw/33s/antiakrasia_reprise/
PS: Here’s a link to a pdf of the chapter “The Intimate Contest for Self-command” from Schelling’s book: http://www.nationalaffairs.com/public_interest/detail/the-intimate-contest-for-self-command I believe that chapter subsumes his seminal 1978 AER article on Egonomics: http://www.jstor.org/pss/1816707
To me it sounds like “Switch to a different drug that’s better for you”.
Referencing long-term consequences could also be viewed as having empathy with ones future self. Instead of thinking “What do I care about the me of tomorrow?”, one creates the impression/illusion of a continuous personality. Maybe empathy for others even evolved piggybacking on empathy for future versions of oneself.
If you are going to manage long term consequences by having empathy with ones future self, there had better be a self threading through time. So I see this as the explanation for why “the self” exists at all.
This is a good response to Blindsight. I’m less sure about it as a response to OB’s version of cryonics, but I need to play taboo with the subject for a while before I can comment further.
I had been playing hold ’em at a casino for about an hour and already turned $40 into $69. A particularly loose cajun on the other side of the table put up an $8 bet. For some reason, I thought of this post and referenced an $8 bet as “about an hour of honest wages.” I folded and was unable to gamble for the next two hands. I called it a night early and left, mediocre winnings in hand.
I think you’re using the term “reference point” differently from how it’s normally used. Normally, a “reference point” is the standard of comparison that you use for judging whether something is good or bad, a lot or a little.
For instance, if you’re used to spending $20 for a shirt, then a $70 shirt will seem expensive, compared to that $20 reference point. But if the $70 shirt was marked down from $150, then maybe it will seem like a good deal, compared to the $150 reference point. Or if you’ve been dieting and you’ve lost 10 pounds in the past month, then you might be happy about that because your old weight is your reference point. But if your goal was to lose 15 pounds in one month, then you fell 5 pounds short of the standard you’ve set and might not be so happy.
What you’re talking about in this post sounds more like construal—the subjective meaning that something has to you. When you think about a McChicken sandwich you’re construing it in a certain way, imagining some aspects of the experience and neglecting others. Your feelings about that sandwich, and your motivations, depend on which parts of it you highlight in your mind and which you don’t.
The official name of the near vs. far mode theory is Construal Level Theory, because near mode thinking involves construing things in a way that focuses on concrete aspects of them (like the taste of that sandwich) while far mode thinking involves construing of things in a more abstract way (like your long-term health). It’s hard to counteract near-mode desires with far-mode thinking, which is part of why it’s so hard to resist temptation. It sounds like you’ve found a near mode, concrete construal of junk food (focusing on the bad heavy feeling you get from eating it) which has helped you to stop eating junk food.
A related idea that I’ve had for a decade or so: when someone (a country, a company, a person) wants to follow in the path of someone else, they shouldn’t aim for their own conception of that other’s location or path; rather, they should aim for what the other was aiming for, at the time they themselves traced their path.
What you are describing is well known in Neuro Linguistic Programming. I’m not super familiar with the terms (yet) but I THINK this is referred to as “chunking”. “Submodalities” and “anchoring” might also be relevant.
More info: http://en.wikipedia.org/wiki/Neuro-linguistic_programming
I think LW could do very well to import a lot of knowledge from NLP and try to see what’s valid and what’s not. I’ve noticed that people here are often reinventing the wheel.
PJ Eby, please chime in on this.
True, but I don’t believe it really has a name there. Tony Robbins refers to them as “references” or “reference experiences”, but I’m not sure if that’s his own coinage or borrowed from NLP.
If you must throw it into a formal NLP term, the blanket term “strategy” is probably what you’re looking for. (The terms you just gave are for things used in strategies, or to modify strategies.)
However, I personally think that this is one area where importing NLP terminology actually obscures more than it makes clear—Robbins’ terminology and the one used here are sufficiently clear to get the point across with a minimum number of entities introduced.
Except when the wheel is, you know, ‘bad’. Lesswrong wouldn’t accept NLP for normative reasons. That said, I would be reluctant to give NLP ‘wheel’ status in any case. “Reinvent the swiss army knife” perhaps. :)