Full text of world-renowned relationship expert John Gottman’s talk titled “The Science of Love” at TEDxVeniceBeach conference.
Listen to the MP3 Audio here:
John Gottman – Relationship expert
So this has been what I’ve dedicated my life to. There are 900,000 divorces in the United States of America every year. Fewer than 10% of them ever talked to anybody about their relationship.
So why would you need a science?
Well, we need a science to develop effective treatment and understanding of how to make love work. Why?
Why should we care about having great relationships?
Well, it turns out that in the past 50 years, a field called social epidemiology has emerged, and it shows that great friendships, great love relationships between lovers and parents and children lead to greater health – mental health as well as physical health – greater wealth, greater resilience, faster recovery from illness, greater longevity – if you want to live 10 to 15 years longer, work on your relationships, not just your exercise – and more successful children as well.
So love has a kind of magic; it’s able to do amazing things. And one of my favorite films is Sleepless in Seattle, with Tom Hanks and Meg Ryan. And I just want to share with you a quote, where Tom Hanks is on the radio with the talk show host Marcia, and she says to him, “Sam, tell me what was so special about your wife?”
And Tom Hanks says, Sam says, “Well, how long is your program? Oh well, it was a million tiny little things. We were supposed to be together and I knew it. I knew it the very first time I touched her. It was like coming home, only to no home I’d ever known. It was just taking her hand to help her out of the car, and, you know, it was like magic.”
And Tom Hanks and Meg Ryan say the word “magic” at the same time. So what about this magic?
Can science really help find the magic of love?
Well, the first step is that we needed a lot of data. We needed to basically understand relationships better, and that’s not something I did alone. In fact, 45 years ago, we built a “love lab,” and this lab was built, in part, by a bromance. My best friend, Bob Levenson, and I created this lab.
And Bob and I became good friends. And we realized that our relationships with women were not going very well; it went from one disaster to another. So basically, two clueless guys got together to build this laboratory.
And then over 30 years ago, a romance with my wife, Dr. Julie Schwartz Gottman. And we decided to work together to see whether we could make a difference.
So basic research was followed by applied research. And we use validated questionnaires, online questionnaires that allow us to assess the strengths of the relationship in the areas that need improvement. And we validated these questionnaires to make sure that we knew what we were measuring and we could measure things reliably and accurately.
So we created this checkup — and over 40,000 couples have taken these questionnaires — and we get a green circle for a strength, a red box for an area that needs improvement.
But we get their story of their relationship. We ask them about how they met. We find out the quality of their friendship and intimacy from this interview. And we collect physiological measures from them as they’re talking to one another: we’re measuring heart rate, blood velocity, skin conductance, respiration, a variety of things like that.
And we score their emotion second by second in this kind of split-screen arrangement, where even though they’re facing each other, we can really code facial expressions, voice tone, nonverbal behavior, and verbal behavior very accurately and reliably.
We have them use a Video-Recall Rating Dial, which they just turned from extremely negative to extremely positive as they’re watching the videotapes, to get their perception of the interaction as well.
And then we synchronize all of that: video, physiology, coded emotion, and perception.
And on the left, you see all the physiological measures we’re collecting, and, you know, a cursor that moves along. And this particular moment, you know, the wife has said something that just makes the husband dissolve in hilarity.
And shared humor turns out to be very powerful in a relationship at reducing physiological arousal.
So, what do we find from this laboratory?
We got over 90% success in predicting either divorce or stability, or the happiness of relationships that were stable. And the major impact of that finding was that Julie and I didn’t get invited out to dinner very often.
But the effects replicated six different times in six studies that Bob Levenson and I did over the whole life course, and that’s probably the most replicated effect in the study of relationships now.
And we’re no longer alone; other labs are getting very similar kinds of results. We followed these couples for as long as 20 years – straight couples, gay and lesbian couples, newlyweds, middle-aged couples, older couples into their late 80s and 90s.
So, what do we find – this over 90% prediction? What predicts?
And we kind of sit down in our lab and talk to a couple about their strengths and areas that need improvement based upon this research.
Well, one thing we do is we create kind of a “Dow Jones Industrial Average” of a conversation. And by doing that, very much like the Dow Jones, it’s really good if the cumulative positive minus negative emotions are going up.
In other words, on the left, we see a low-risk couple, where, you know, there really is much more positive than negative, in general, right? They go up and down.
And on the right, the high-risk couple, where basically, they go down all along their interaction. Well, turns out that one thing we discovered was that the way the conversation starts in the first three minutes of a conflict discussion will predict, 96% of the time, whether they are a low-risk couple or a high-risk couple.
So startup – the way each person starts before they start influencing one another – is very critical in this prediction.
The other thing we learned is something we call the Roach Hotel Model of Relationships. Remember the Roach Hotel? The roaches check in, but they don’t check out. That’s a really good roach trap.
Well, negative emotions – anger, sadness, disappointment, fear, all of these emotions – the negative emotions for unhappy couples become like that roach hotel: once they check in, they don’t check out; it’s hard to exit and easy to enter this negative affect state.
And it turned out the balance of positive and negative emotions is our key index of this magic that Tom Hanks is talking about in Sleepless in Seattle — the balance of positive and negative emotions.
So this ratio of positivity to negativity during conflict in unhappy couples – that’s why this slide is in red – was 0.8:1, just a little bit more negativity than positivity.
And I was completely wrong; I thought that would be a great relationship. You know, in my relationships, you know, that had failed, we were much more negative than positive, so I thought if it was balanced, it would be great.
But no, it turned out that, actually, the balance of positive and negative emotions during conflict in relationships that were stable and happy was five to one. There was five times as much affection, humor, interest in one another, excitement, connection than there was hostility, disappointment, anger, negativity. So there really was a balance that was way balanced toward positive emotions in happy, stable relationships.
So this five-to-one positive to negative ratio has become pretty widely known. I left a Starbucks in Seattle recently, and a guy drove by in his pickup truck, and he rolled down the window and he said, “Hey, five to one, right?”
I said, “Right.”
So we both gave us the thumbs-up sign. So that’s our index of the magic.
But then we thought, “OK, you have an index, but how do you change that index? What do you need to really make a difference?” And so our hypothesis was that you needed three things: physiological calm and trust and commitment.
So what about those things? Can we actually measure those in our laboratory? And can we actually make a difference?
And this is right. Well, the first one, physiological calm, was kind of a balance in physiology. The couples whose heart rates were lower, whose blood was flowing less rapidly, who weren’t sweaty as much, those people actually seemed kind of boring, but they had great relationships. They were gentle with one another, not hostile, they reassured one another – very interesting balance in physiology created by their behavior.
And by the way, we found we have to measure these things. So I have a pulse oximeter right here, and it allows me, when I press the button, to really find out what my heart rate is and what the percentage of oxygen is. And right now, my heart rate is 110 beats a minute, and my percentage of oxygen is 92%.
So that sucks. I’m highly aroused physiologically.
So you can’t tell what’s going on physiologically with somebody by looking at them; you have to actually measure it. And we have to build physiological calm in a couple.
Now, why? Why is that so important?
Well, it turns out that when people are calm, they can take in information, they can listen, they can be empathetic, they have access to their sense of humor. That’s very important.
But when they’re flooded, when there’s a diffuse activation of various parts of the autonomic nervous system, they’re much more likely to be in attack-or-defend mode, right?
So if a therapist doesn’t know what’s going on physiologically and says, “Can you summarize what your wife just said? Summarize what your husband just says? Can you validate and be empathetic?” They can’t do it if they’re physiologically aroused, right?
So the therapist actually has to know that to help them calm down. So physiological calm is good. That’s what we learned.
So, part of that trio of variables, we know we have validation that that’s important. But the magic also comes, we hypothesize, from building trust.
So what’s trust?
It’s something that people talk about, but can you measure it? And it turns out we can measure it reliably and validly. And a trusting relationship, a mutually trusting relationship, really leads to intimacy and great sex; a distrusting relationship leads to loneliness.
And guess what? The major reason people have affairs is not because of sex or desire; it’s because they’re lonely. And they have found somebody who finds them interesting and wants to talk to them in the context of a relationship that really isn’t working very well.
So what is trust? How do we define it? How do we measure it?
And here we use mathematics – the mathematics of game theory. And we can actually measure this idea that mutual trust comes from when both partners are maximizing the benefits of both people, not just one person’s benefit against the other person.
So I’m always thinking about “how does my wife see things?” I can walk into a kitchen now, after 30 years of marriage, and view that kitchen the way my wife would do it. So I can say, “Ah, Julie would be upset by that,” so I’ll clean it up, and, well, now I know when she comes in, she’s not going to notice anything that’s really disgusting.
You know, I’m more of a slob. So, you know, for me, it was fine, you know. But, you know, people who develop mutual trust, they really always have their partners in their heads, not just themselves, and they’re thinking about both of them.
But the magic also comes from building commitment. So what does this mean?
Well, we’ve learned a lot about commitment from the systematic research of a woman named Carol Rusboldt and another woman named Shirley Glass.
And what we’ve learned is that in relationships, there is a turning point, a key turning point. When things aren’t going well, when your partner is hostile or irritable or emotionally distant, if you really, in your own head, are cherishing your partner and nurturing gratitude for what you have and saying, “This is my journey. Julie is the love of my life. And I’m damn lucky to have this person in my life,” that turning point leads to loyalty.
On the other hand, betrayal comes from, really, at that point when things aren’t going well, making negative comparisons between your partner and real or imagined alternatives.
Betrayal leads to dissolution very reliably, whereas commitment leads to loyalty. And then, when you have that sense of commitment in a relationship – both of you have that – you really have a safe place, you have that magic.
And with over 90% prediction, with these basic variables, the mechanism by which relationships work – trust, physiological calm, and commitment – can we actually create the mathematics of love?
Now, why would you want to create mathematics in love? Why do you want to put math and love together?
Some people think that’s really disgusting. Well, if you can create a mathematics, you really will understand not just the static nature of commitment, trust, and physiological calm affecting a relationship and creating the magic, you actually will understand the dynamics of the interaction as it unfolds over time.
And that’s what therapists are working with, right? In their offices, they are looking at how people interact, and they want to influence the dynamic. So we need a dynamic analysis of data over time.
What pushes the interaction in one direction versus another?
So, because we get data over time, we get behavior, perception, and physiology all integrated over time, very much like that Dow Jones Industrial Average, we can create the Gottman-Murray love equations.
And so, James Murray, who is the father of mathematical biology, and I worked together. And because we’re not very intelligent, it took us 15 years to get these equations.
But when you look at them, you’ll see they’re obvious. So what we’re trying to do is predict what happens at time “t+1” in this graph of the husband’s interaction and the wife’s interaction. And by the way, this holds for same-sex couples as well; we say wife and husband only because it’s a little bit more convenient to talk about it.
But can we go for what’s happening in this graph of behavior or physiology or perception at one time and predict at a later time?
And in fact, we can develop these equations. So on the left of the equal sign, we see the wife at time “t+1” in the equation above, and on the bottom, the husband’s.
So how do we predict what’s going to happen next?
That’s what we want to know. That’s what the therapist wants to know.
Well, we have one parameter, and it’s denoted A and B, and that’s the old startup, right?
How they start the conversation before they start influencing one another is really critical. And so that’s the theory that the way they start up, the way the wife starts up, that one constant is very, very important in predicting what’s going to happen next.
But then there’s another parameter, which is R1 and R2, and that predicts how much emotional inertia a person has. So some people, when they get angry, boy, they stay angry a long time. It’s hard to move them. They’re kind of like a Mack truck going downhill. And in physics, that’s called inertia, right? It takes a lot of energy to move them off course.
Other people are more like a feather in the wind; they have low inertia, right? Well, they have no stability either.
So what’s the right inertia?
Well, we don’t know, but that’s another parameter – their emotional inertia. That affects how much influence the partner can have, which is our third parameter: the influence function.
So in the wife’s equation, there’s a husband, and in the husband’s equation, there’s a wife. Well, we can actually specify with a theory what that influence function might look like.
So here’s a husband along the horizontal axis, and the vertical axis is his influence on the wife. So, there are some points, but we need a theory that helps us find the best fit to the data, right?
So, our theory really said, well, maybe it looks like this: There’s one part of it when he’s being positive, he’s being nice to her, what’s the impact of that? The steepness of the slope tells us what the impact of positive emotions are, like affection, humor.
And then the other one is how influential is he with negativity, with anger, disappointment, with hurt feelings, with sadness. And that slope gives us our idea of the influence that positive and negative emotions have.
But then we thought, well, maybe when they get to a certain threshold – that we have denoted by K sub R – maybe what happens is they can repair.
And so, that other part of the curve tells us when does repair cut in and how effective it is. And what we discover is that if you repair early in the conversation, you’re much more effective than if you wait. OK, that was kind of a surprising finding.
And on the other hand, what about positive affect? Maybe you could amplify or diminish positive affect. And it turns out, yes, you can. By turning toward your partner when your partner needs your attention or interest, the positive affect can be either by turning away, diminished, or by turning toward, amplified.
So these are the parameters in our theory. But now we do a very cool thing. We actually can create a portrait of the couple’s interaction, a dynamic portrait of how it unfolds over time.
And what we do is we put the husband’s positive and negative behavior on the horizontal axis and the wife’s on the vertical axis. Going up is positive; going down is negative. That gives us quadrants one, two, three, and four.
And we can describe it by two bunnies on the beach in quadrant labeled one and two bunnies in the storm in quadrant three. So we really want our couples to be two bunnies on the beach rather than two bunnies in the storm.
So this gives us a new goal of couples therapy, right?
All right, so this phase space portrait becomes very clear. This quadrant one is when they’re both positive, and we have another quadrant where they’re both negative. So we’d like them to really not spend much time in that quadrant, or these quadrants where one of them is positive and the other is negative.
Although it might be good to be there because maybe one counterbalances the other, so we don’t know yet. So we have those quadrants to look at as well.
OK. There’s our relationship phase space. Now, let’s look at an actual conversation and see how it unfolds. And there’s a dot moving as the couple talks to one another; this is a real conflict conversation a couple is having. And look at that dot moving all over those quadrants in phase space.
So now we see there’s a dynamic unfolding, but it’s kind of hard to see that thing jumping all over the place and know what’s good and what’s bad. We need equations, and we need to know what’s making that dot move. What’s actually compelling that dot to move?
Is there some attractive force, kind of like gravitation? Is there an attractive force that has, maybe, the power of Jupiter to pull them toward the negative attractor? In which case, the relationship will be in a lot of trouble.
Or is there a positive attractor that puts them in quadrant one that is powerful? Or is it weak, and the therapist really needs to help make that much more powerful?
So, here’s our old friend, the startup. Now, this couple is starting in a positive place, right? They’re in quadrant number one, the positive-positive quadrant. And there’s a choice here. They either move in that direction, toward that star, that blue star, which is the attractor in the phase space that pulls their interaction toward a positive place, especially if it’s strong, or they might move the other way.
They might actually cycle around another attractor that’s there that moves them in the negative-negative quadrant. And that’s our bunnies in the storm.
Now, we don’t want them there, or we don’t want them to spend very much occupation time in that place. And the attractors are the forces that make the dot move. So we’re looking at not just a graph and face space, but we’re looking at a flow diagram, almost like a fluid flowing.
And you’ve seen this kind of phase space diagram with hurricanes. Hurricane Irma, this comes from that hurricane, where the dot is moving, right? And there’s a force moving it toward an attractor up there. And we can actually do the same thing with a couple’s interaction.
So, OK, that’s a theory for getting a dynamical picture. Does it work?
Well, this is what a stable, unhappy marriage looks like. Look at all the flows, look at the arrows, and what are they flowing toward? A negative attractor, for our bunnies in the storm, right? We don’t like that, you know.
And well, what about divorced couples?
Well, it turns out, with divorced couples, you get this picture. There are three attractors in that negative-negative quadrant, and one is unstable – the one between the two stars and the two stars are unstable – so what we get there is turbulence. Not only is their flow toward the negative-negative quadrant, but a lot of energy is expended as they move from one attractor to the other.
A turbulence is probably what predicts divorce versus being in a stable, unhappy marriage. Now, I don’t know which is worse, but they are different.
But the real test is, what about good relationships – happy, stable relationships? How do they look? And it turns out when you put this together with real data, you get this picture. So yes, it’s true.
Happy, stable relationships have a positive attractor – very, very simple phase space diagram. And no matter where they start the conversation, where that startup is, they’ll go toward a positive place, which is kind of amazing. That’s the dynamics of a good relationship.
So we did it. We actually were able to describe, mathematically, the dynamics.
Now, here’s another thing you can do with the math. With the math, we can simulate any couple under imagined conditions; in other words, we can do thought experiments. We can do experiments by changing the parameters, and mathematically, we can select the best intervention to help a couple just using math.
So for example, here’s a couple. And what you see on the left are these sliders where we can change every parameter in the math equation, right? We can change their startup, we can change their emotional inertia, we can change their repair threshold – how much they repair, when they repair.
And here we see, you know, the green lines that show us the influence functions and the arrows that show us the phase space portrait, the flow. So let’s take a look at an actual couple.
And here we see this couple are two bunnies in the storm; they have that negative attractor. And we can actually move the sliders to say, “What would they look like if their startup was just more positive?”
So we move – and this is a thought experiment – we move the couple to a more positive place by changing those parameters at startup, and then we take a look at what the result is of the math?
And look, there are two bunnies on the beach. They’re in that positive-positive quadrant. And then all we need is a technology for changing startup, which we have. So we’ve created the technology for changing these parameters, and now the therapist can change their startups. And then they get the magic.
Research findings are that the magic does require this powerful trio of physiological calm, trust, and commitment, and we know how to build it. So, science did it.
When the magic is there, we find that we get that five-to-one ratio. Now we can understand the magic because with a positive attractor, no matter where the conversation begins, it would be drawn to the positive, positive quadrant. We’ll get the five-to-one ratio – Yahoo! there’s the magic, OK?
So, just to summarize, what we’ve really said is love relationships are important. We can assess their strengths and challenges scientifically. There is magic in love that lasts forever – we know that. And the five-to-one ratio indexes it; we have strong prediction. But we can mathematically model the magic.
The magic requires calm, trust, and commitment. And maybe we can help avoid disaster and help couples find their magic, which is turning out to be true in randomized clinical trials.
So perhaps now, the magic of great love is a bit less of a mystery.