« China dam of the day | Main | The culture that is French, a continuing series »
The symmetry thesis
The thesis is simple, and almost everyone disagrees with it upon first hearing.
The symmetry thesis: A given person likes (loves) you as much as you like (love) him or her.
I have encountered many apparent refutations of the symmetry thesis, but with time most have turned out to be spurious. I find the symmetry thesis a surprisingly strong predictor of human behavior and inclination.
Do I want to know how much you like me? It is simple. I imagine how much I like you. (If you do the same, are we circular? Or does some kind of fixed point theorem apply?)
Let me rule out or explain some obvious "counterexamples." If a guy
stalks you, and you can't stand him, the reality is that he is probably
more hostile to you than loving. The thesis fits.
Break-ups are tricky and they provide the best counterexamples. But who really left whom is not always obvious; it can take several years to figure out what was going on. Often the leaving party is the one who first develops a narrative of how things might be different; this is distinct from liking or loving the other person less. Other people leave pre-emptively.
Unilateral crushes are possible and indeed common, although with repeated contact they usually collapse into symmetry, one way or the other.
I can imagine several (non-exclusive) mechanisms in support of the symmetry thesis. Perhaps "having a connection" -- which is mutual by nature -- is the key to true liking and attraction. That is my favored view. Note that it creates a possible exception for people who can like or love others without having any real connection with them. I tend to think of such likes as delusional.
Alternatively, perhaps at least one person is a "fraidy cat," and
won't let himself or herself fall for the other, or even like the
other, without witnessing signs of reciprocity. The two people then
lead each other down the pathway of like, in a kind of low-key
intertemporal seduction, sans the sex. Or with it.
Perhaps we like other people for their intrinsic qualities less than we pretend. Mostly we like people for liking (loving) us.
Yes I know that most of you don't believe it, and have plenty of counterexamples to offer. But keep it in the back of your mind, and see if it proves useful over the next few years.
Posted by Tyler Cowen on May 23, 2006 at 05:56 AM in Science | Permalink
Comments
I think a common theoretical arguement against the symmetry thesis is the "Groucho Marx Syndrome", made manifest in Woodie Allen's Annie Hall: 'I refuse to join any club that will have me as a member.'
This maps into the truism 'wanting is better than having,' which has moderate support from hedonic theory.
Of course, after Alfie rejects Annie, he regrets his decision and his love returns.
And certainly, the exhortation to "love thine enemy" has immense practical value if the thesis holds.
Posted by: Allan Friedman at May 23, 2006 6:51:13 AM
The obvious counterexample would be our relationship toward celebrities. I may love Jessica Alba and loathe Tom Cruise, but these feelings are surely not reciprocal, because neither of them knows me.
Posted by: The Other Brock at May 23, 2006 6:54:13 AM
Nor does The Other Brock really 'know' Jessica or Tom. It seems that Tyler is talking about relationships with some interactive component. I think the theorem is most useful in the 'long run' of relationships, that an equilibrium is reached where both parties feel similarly affection for each other.
Posted by: Robbie Allan at May 23, 2006 7:39:28 AM
But that's not Tyler's thesis. Perhaps what he really means is:
1. "The expected value of Person A's like/love for Person B is equal to Person B's like/love for Person A"
or
2. "Over time and repeated interaction, Person A's like/love for Person B asymptotically approaches Person B's like/love for Person A and vice versa."
But his version of the symmetry thesis is much stronger than either statement. I don't think I agree with Cowen's thesis as stated—too many counterexamples. But the commentary about "collapsing" suggests that TC really means the weaker statement (2). I'm not sure that holds, either, except in the Galbraithian sense that in the long run we're all dead.
Posted by: Ted at May 23, 2006 8:10:21 AM
Wow, I really like this idea. IMHO the issues with stalkers, crushes, and celebrities are resolved if you make the traditional distinction between lust and love. No, you don't like or even really know the object of your crush. Neediness isn't like either, which is one reason it is repellent.
The biggest problem with the symmetry thesis, though, is in explaininng parent/child and parent/teenager relationships. The "teenagers are emotionally unstable and constantly shifting" thesis may be more accurate there.
Posted by: DK at May 23, 2006 8:28:12 AM
Ted has it--I think the thesis should be restated as "Over time, a given person's like (love) toward you will trend to the same level that you like (love) him or her.
Posted by: Sandy Smith at May 23, 2006 9:21:19 AM
I really dig this theory.
I think it may also help to explain why insecure people do things (consciously or not) to push others away from them: they make the other person end up feeling as ambivalent about them as they feel about deserving the relationship in the first place.
Posted by: Jared at May 23, 2006 9:57:48 AM
One of the odd things about Proust for me was how he so fervently pushes the opposite of TC's thesis: that every relationship has a powerful, frequently unstable and reversing, asymmetry, and that this is the tragedy of love for humans. I don't find either Proust or TC persuasive; I think each relationship has its own dynamic.
Posted by: Grant at May 23, 2006 10:33:51 AM
So much for Plato (the Phaedrus in particular) and centuries of the lover/beloved dynamic.
Posted by: Anderson at May 23, 2006 10:50:44 AM
Tyler, your explanation of the stalker counterexample sounds a bit like the "No True Scotsman" fallacy. For any counterexample to your thesis, you're pretty much free to remove it from the analysis by saying "That's not actually love, that's lust/power/enfatuation/special case X", and maintain that the thesis is true and unrefuted.
I agree that the thesis *seems* to be true, because there's something sensible behind the fallacy: I'm reluctant to allow the word "love" to apply to something lacking reciprocation.
You can fix this if you can come up with (and stick to!) a good definition of "love"---a very old problem. I recall a good one in Freud's "Civilization and its Discontents" (!!): love, he said, was a delusion in which you are unable to distinguish your partner's happiness from your own.
Posted by: Ben M at May 23, 2006 11:13:05 AM
I think this is somewhat true but much weaker than stated. People have a powerful desire to be liked if someone doesn't like a person a person will create a negative story about that person to explain why that person's opinion doesn't really matter. Conversly if Person A likes Person B, Person B will tend to form a positve story about Person A in order to inflate the worth of their attention. Of course both actors are playing by these compatable sets of rules so there is a tendancy to create powerful feedback loops.
Posted by: Michael Foody at May 23, 2006 12:01:44 PM
I believe that it is important to distinguish between attraction (or, call it love, fondness, affection, etc.) and the value a person places in a relationship.
A relationship has costs (compromises). Relationships only work when both sides value the relationship more than the costs that are required to make the relationship work. I agree that, over time in long term relationships, the values that two parties place in a relationship tend to converge.
In most discussions about love, I find that 1) the value a person places on a relationship and 2) affection are used interchangeably. Call me cynical, but I do not believe that this is the way people actually work. I think that if you consider these elements separately, it is much easier to see where the symmetry theory does and does not apply.
Posted by: ipninja at May 23, 2006 12:06:09 PM
I think Tyler's insight is generally correct and we can math this thing up a bit to get something real.
Think of a pair of difference equations
Delta (Y's Love for X) = alpha*E[(X's Love for Y)]
Delta (X's Love for Y) = alpha*E[(Y's Love for X)]
The intial points are determined by X and Y specific charateristics.
This implies, that if the intial points are close in feelings that we will move quickly towards some equilibrium. However, if they are far away there could be some interesting chase dynamics. Especially if we move from multiplication by alpha to a more complex mapping. Heavy love by one party could overcome antipathy from the other.
What would also be fascinating is if there is some exclusivity to love. This would imply that two people alone might very quickly graviate towards love but in an envrionment with multiple possible partners chaotic dynamics become quickly possible.
Also, the expectation operator is important because it helps explain love from afar. It can persist because there is not enough information to fully update the expectations operator. You will also note that crushes tend to increase if the person gains information which causes them to expect that the other person "would love them" if they got to know them.
Few people, I believe, have crushes on those who they anticipate would hate them even under the best circumstances.
Posted by: Karl Smith at May 23, 2006 12:48:19 PM
I would think that celebrity is the best counter-example.
Posted by: Eric Anderson at May 23, 2006 1:05:15 PM
Yeah, this feels more like a statement about the long run rather than the short run. As soon as one (or both) party recognizes the imbalance in feelings, a cycle of actions takes place to reconcile the imbalance. If your Marginal Love is greater than the other's, you "devalue" the person in some way (e.g., find faults that devalue your prior assessments, avoid them, healthily accept that you're just not meant to be together (ha!)); or, if you are the beloved, you "devalue" yourself or the relationship (e.g., don't return phone calls, fewer visits, stop smiling as much during interactions).
Of course it's imperfect: rationing access to the beloved may incite the lover that much more (i.e., stalkers, hopeless/helpless romantics). Not that I speak from experience....
Posted by: RSaunders at May 23, 2006 1:14:03 PM
I think Ben strikes on a real slippery issue. A firm and established definition of "love" needs to be established for any meaningul discussion to take place. I think that the Thesis does however have some explainatory power, based simply on observed experience, and personal experience, I am somewhat driven to accept the thesis, though I could simply be too inexperienced to have a sizable sample size.
The other thing that I think that the Thesis, at least in my observations as a college student, definitely applies to a more sexual interaction between individuals. It's my experience and observation that a couple that seems to be more sexually motivated, seems to hold to the thesis in the opposite direction from love and instead of one of mutual ... i think gratification is the tactful term I'm looking for..
Just an observation from a humble college Econ major.
Posted by: AnonymousOne at May 23, 2006 1:37:49 PM
In every relationship, there is a moment of maximum symmetry. Sadly, a man's emotional high points will rarely coincide with those of the woman he loves. Women and men are just too different. The longer a relationship goes on, the more moments occur when there's no match at all. And perhaps the mutual high points are all pretty much the same. The mismatches are all different, and so more memorable.
Posted by: Robert Speirs at May 23, 2006 1:54:11 PM
Reminds me of a variant of the tit-for-tat strategy in a multi-period prisoner's dilemma game with like/like and dislike/dislike as stable solutions.
Given common preference structures, spurned attention investments result in a redirection of one's affection. Don Juan is the classic counterexample: He directs his attention elsewhere after the "kill".
The fun starts with multi-person constellations, opportunity costs and social feedback mechanisms.
Posted by: jaywalker at May 23, 2006 1:57:13 PM
If the symmetry thesis is correct, then mathematics without meaningful units must think I'm silly.
Posted by: James at May 23, 2006 2:06:17 PM
I love him and he doesn't love back is the oldest, most poignant (not to mention easiest to relate to) stories of all time. Also, the author of "He's Just Not that In To You" definitely would take exception to this thesis. I can't come up with compelling non-anecdotal evidence but I know in my stomach that this is wrong.
Posted by: Sarah at May 23, 2006 2:18:17 PM
Ben M. nailed it. Tyler's analysis goes down in flames and in a just world this thread would be over.
Of course, the world is not just, love is not symmetric, and the thread will continue, so I just want to point to parent/child love, which is frequently and obviously asymmetric (examples left as an exercise left to the reader).
Posted by: Cog at May 23, 2006 2:50:40 PM
I'm wondering; is this thesis meant only to apply to positive feelings, or can we describe like on the continuum from hatred to adoration? If we can use the continuous approach, then I think that the thesis isn't very good. It seems that many people have taken a good look at the positive side of the spectrum. Let's look at the negative side.
I imagine that there are many individuals who merely tolerate my existence that I think of as genuine friends (I recognize that I am, in fact, an a**hole, so many people I think of as friendly are probably just being polite). For that matter, I can have a long-seething hatred for an individual (something akin to "The Cask of Amontillado (sp?)") who has almost no idea that I wish him or her ill. The individual may come to know this, but only too late. Is that sufficiently long run?
[That said, I really want this thesis to hold true, particularly on the positive end. It sounds hopeful (e.g., I love her, and if I keep loving her, someday she'll love me too!). So thanks, Tyler, for the happy thought!]
Posted by: hamilton at May 23, 2006 3:43:07 PM
Without some kind of constraint on initial conditions (I prefer brunettes or bookish women or whatever) this theory would generate infinite equilibria, right (there wouldn't be a fixed point in the system)? I think it describes dynamics fairly well (if I have unrequited love for someone either I will eventually get tired of being frustrated and move on or she will come to love me) but doesn't give any predictions about which people are most likely to end up in a couple together.
I guess the real question is if most people seek out someone who they think is likely to develop feelings for them. Couples would then be formed by mutually-reinforcing self-delusion.
Posted by: Mark at May 23, 2006 5:34:15 PM
I have stalked many women, and loved them all.
Posted by: ricardo at May 23, 2006 9:52:17 PM
This is off topic but I wanted to respond to Ted about the "Galbraithian in the long run we are all dead". That was stated by Keynes long before Galbraith. Keynes was asked about long term investing and Keynes said in the long run we are dead.
Posted by: Murphy at May 24, 2006 11:15:21 AM
