The Natural Principles of Love: The Natural Principles of Love

John Gottman and Julie Gottman Gottman Institute

The Natural Principles of Love

In this Original Voices article we summarize the past four and a half decades of our work on relationship stability and happiness and explore the theoretical implications of that empirical research. First, we briefly review the laboratory research, clinical work, and the mathematics used to understand our results and build our theory. Then we describe the sound relationship house theory, constructive blueprints for managing conflict, and the three phases of love. We use the term love in the narrow sense of the primary emotions that draw people together to form a lasting, committed relationship between lovers, regardless of sexual orientation. Although we began with no theory at all, we were led by our data and our clinical work at the Gottman Institute.

In 2005, 14 scholars got together to review what the benefits of marriage might be, as determined by social science research (Wilcox, Doherty, Glenn, & Waite, 2005). Their report was cautious and quite lengthy. They wrote that it was happy marriage itself that predicted very positive life outcomes for men, women, and their children. That report was a resounding endorsement of marriage, and yet these conclusions were only part of the story. The results were a small part of a much larger scientific literature linking the quality of people's closest relationships to health, longevity, and well-being. Forty years ago, these findings initially surprised the epidemiologists Leonard Syme and Lisa Berkman, but they have held up over time, and they have

Gottman Institute, Seattle, WA (jgottman@).

Key Words: Conflict, emotions, game theory, interaction, love, relationships.

created a vital new field called social epidemiology (see Berkman, Kawachi, & Glymour, 2014). Ignoring what is cause and what is effect, there is no doubt that people in happy, stable, committed relationships--versus people who are alone, in uncommitted relationships, or in unhappy or unstable relationships--live significantly longer, are healthier physically and psychologically, become wealthier, and have children who do better in most aspects of living. Therefore, there is no question that we can precisely identify successful and unsuccessful relationships, and measure the effects of both. Relationship success or failure has enormous consequences for people everywhere on the planet.

We start with the aforementioned history to contextualize our work against so-called poststructural theories of relationships, such as strategic and narrative therapies that reject the idea that some relationships fail and others succeed. Their view is that every relationship is unique and whatever happens is just fine, that it is absurd to talk about success and failure, that everything is relative, and that culture and perception determine everything (see Gurman, Lebow, & Snyder, 2015). These theories lead to therapies that apparently never fail, because they consider every possible outcome fully acceptable. These theories also view a scientific approach to love as if its goals were to paint the entire world gray and claim that everyone is the same everywhere. Our work is based on the opposite premise: Relationships do indeed fail, and that outcome is not what couples hoped for at their wedding or commitment ceremonies. Relationships fail at a great cost to everyone. That is not to conclude that divorce is always unwelcome or that divorce needs to be a lifelong tragedy.

Journal of Family Theory & Review 9 (March 2017): 7?26

7

DOI:10.1111/jftr.12182

8

Journal of Family Theory & Review

Why Care So Much About Understanding Love?

Can science bring clarity where artists have tried so hard and failed? Is there wisdom to be learned at all? Do empirical findings hold? Do they replicate? Can we understand our results? Can we discover truths that may hold everywhere on our planet? After four and a half decades of research on relationship stability and happiness, we believe that the answer to these questions is yes. This article is about our understanding of what makes relationships long lasting and happy. We use the term love in the narrow sense, to mean the primary emotions that draw people together to form a lasting, committed relationship between lovers, regardless of sexual orientation.

Data, not theory, are what we brought to this work. In this article we summarize what we have learned through empirical research and therapy on relationships and explore the theoretical implications of that research. We have achieved high levels of prediction of the future of heterosexual and same-sex relationships. We have studied relationships across the entire life course and have been able to predict successful life transitions, such as to parenthood and to retirement. We have also applied these methods successfully to the study of parent?child relationships (Gottman, Katz, & Hooven, 2013; Havighurst, Wilson, Harley, & Prior, 2009). For most of these years John has collaborated with Robert Levenson in basic longitudinal research about relationships. For the past 20 years we have collaborated as a husband?wife team, in clinical work, and in randomized clinical trials that reveal the natural principles derived from basic research and show that these principles lead to interventions that are successful at preventing relationship distress during the major transition to becoming parents (Shapiro & Gottman, 2005) and at healing most ailing relationships (Babcock, Gottman, Ryan, & Gottman, 2013), including very difficult relationship problems such as situational domestic violence (Bradley & Gottman, 2012).

We have also been successful in understanding our predictions and building our theory, using laboratory research, clinical work, and mathematics to understand our results. Mathematics has played a large role in our work. We have employed mathematics in many phases of our work: (a) in finding stable sequential patterns observed in couples' interaction, using the mathematics of information theory and

time-series analysis, with Jim Sackett, Roger Bakeman, and James Ringland (see Bakeman & Gottman, 1986/1997; Bakeman & Quera, 2011; Gottman, 1979, 1981; Gottman & Ringland, 1981; Gottman & Roy, 1990); (b) in studying trust and betrayal, using the mathematics of game theory (Gottman, 2002); and (c) in revealing the complex dynamics of interaction using the mathematics of nonlinear differential equations, with the mathematical biologist James Murray and his students (Gottman, 2011, 2015; Gottman, Murray, Swanson, Tyson, & Swanson, 2002). As a wife?husband team we have combined sensitive and intense clinical work--led by Julie--with subsequent randomized clinical trials to test cause?effect relationships, to prevent relationship disasters, and to try to help ailing couples.

This is supposed to be a theory article, but when we began this research in the 1970s--contrary to what we all learned in graduate school--Robert Levenson and John adhered to no theory and had absolutely no hypotheses. Their goal was simply to observe, describe, measure, and find patterns that replicated over studies. Robert and John were not limited to studying behavior. They were not behaviorists, nor were they psychoanalysts, nor were they object-relations theorists, nor were they structural theorists, nor were they existentialist theorists, nor were they attachment theorists, nor were they narrative theorists, nor were they solution-focused theorists, nor were they strategic therapists, nor were they systems theorists. In fact, they were not theorists at all. Mostly, they were dust-bowl empiricists studying the role of emotion in relationships. Their first task was simply to describe, looking for a convergence among multiple methods. They included self-reports of experience (through interviews and questionnaires), observed interactive behavior with cameras and computer-assisted observational coding, assessed human physiology, and used video-recall ratings with synchronization to the video time code. They wanted to measure all parts of emotion--behavior, perception, and physiology--all synced together with real-time couples' interaction. Of course, we weren't entirely without ideas of what to study; our work was definitely influenced by the context in which we worked, especially by the development of psychophysiology (e.g., Obrist, 1981), the general systems therapists (e.g., Bateson,

Natural Principles of Love

9

Jackson, Haley, & Weakland, 1956), and the quantitative study of emotions in the human face, particularly by our colleague Paul Ekman (e.g., Ekman, 2015).

Methods

Robert and John designed a lab that synchronized the video time code to physiological measures and to a rating dial that people turned from "Very Negative" to "Very Positive," revealing their perception of their interaction, and we had a computer that did this job of synchronization. We wanted to get samples representative of the demographics of the city where we worked. We also wanted about equal numbers of happy and unhappy couples in our studies, so we oversampled these groups. We learned that couples' interactions over time had as much as 80% stability, and we saw that there were both "masters" and "disasters" of relationships. We have fairly low standards. A "master" couple is stable, and both partners are at or above the mean of 100.0 in relationship satisfaction on the Locke-Wallace (1959) or the Dyadic Adjustment Scale (Spanier, 1976). By a "disaster" couple we mean a couple who either breaks up or stays together unhappily (at least one partner is at or below 85.0, which is one standard deviation below the mean of 100.0). Couples came into this lab after having been apart for at least 8 hours and then talked about their day. They filled out questionnaires measuring their relationship happiness. They talked about their day after we attached the sensors measuring heart rate, respiration, blood velocities to the ear and finger of their nondominant hand, the amount they were sweating from their palms of their hands, and how much they jiggled around in their chairs, and after we had obtained a 5-minute silent baseline. There were two cameras in the lab, each giving us a full-face picture of a partner, and they were electronically merged into one split-screen picture with a running time code. After the 15-minute conversation about the events of the day, they were interviewed about what they argued about and asked to try to resolve the major hot issue in their marriage in the next 15 minutes, after another 5-minute baseline. Then they chose a topic from a list of positive topics to discuss for 15 minutes. They had that positive topic conversation after another 5-minute silent baseline. Then, in another appointment, they viewed their videotape and turned the rating dial, also while hooked

up to the physiological sensors. That was the experiment. We did nothing to help them. After 3 years we recontacted the couples and they again filled out questionnaires measuring their marital happiness.

Our video-recall rating dial has proved itself, over the years, to be quite valid. It is a good predictor of the future of a relationship, and it gives us a window into the world of perception. John's postdoc William Griffin (2002) applied the technique of looking for sequential patterns using a method called "hidden Markov analysis" and demonstrated its validity by differentiating happily from unhappily married couples. In another study by Robert Levenson and his student Anna Ruef (Levenson & Ruef, 1992), they had couples use the rating dial twice, once to indicate how each person felt during the interaction and a second time to try to guess how their partner felt during the interaction. They discovered that people were accurate at guessing how their partner felt to the extent that they relived their partner's original physiology during the interaction as they watched the video. Therefore, using the rating dial, they discovered what they called "a physiological substrate" for empathy. In our 20-year longitudinal study, the rating dial, coupled with behavioral coding of emotion, could even predict which husbands would die early, if their marital conflict interaction resembled a competitive, win-or-lose zero-sum game, and which would live longer, if their interaction were a more like a cooperative win?win game (Gottman, 2011). In Levenson's lab, Haase, Holley, Bloch, Verstaen, and Levenson (in press) could even predict which type of chronic physical illnesses people would develop from their specific emotional behaviors during marital conflict 20 years prior; angry people developed chronic cardiac illness, and stonewalling people developed chronic musculoskeletal illness.

Predicting the Future of a Relationship

In the early 1970s psychology was actually at somewhat of an impasse. Walter Mischel (1968) wrote an important, challenging book noting that personality psychology had done a poor job of understanding and predicting human behavior, because even the best measures were able to reduce only about 9% of the uncertainty in prediction. Mischel said that was unacceptable. Therefore, after 3 years, as we followed up with our first 30 couples, we were amazed that

10

Journal of Family Theory & Review

we could account for about 80% of the uncertainty in how their marital happiness changed over a 3-year period, even controlling for initial levels. Furthermore, the results were clear. For example, the couples who became unhappier over 3 years were initially significantly more physiologically aroused than the couples that eventually became happier. Their hearts beat faster, their blood flowed faster, they sweat more from their palms, they jiggled around more, they rated their emotions as more negative on the rating dial, and they were far more hostile (more criticism, defensiveness, contempt, and stonewalling) when discussing the events of their day, a conflict, or even a positive topic than were the couples who became happier over time. Now we had some hypotheses.

Over the next 23 years, as we did that study over and over again, across the whole life course, following couples for many years, we replicated and extended these findings. We also spent a dozen years studying committed gay and lesbian couples. We studied couples going through major life transitions, primarily the transition to becoming parents, and--for 20 years--the transition to retirement and old age.

In 1987, at the University of Washington we built an apartment lab that was designed to be like a bed-and-breakfast getaway. It was on the beautiful Montlake Cut of the medical school campus, overlooking a park, with boats traveling between the salt water of the ocean and the freshwater Lake Washington. A full 130 newlywed couples, each just a few months after their wedding, arrived at 9 a.m., usually on a Sunday, to spend 24 hours in the apartment lab (called "the love lab" by the BBC). The only difference between this lab and a usual bed-and-breakfast getaway was that we had three cameras bolted to the walls to follow all their movements, they wore Holter monitors to track two channels of electrocardiogram, when they urinated we took a sample to measure stress hormones, we took blood from them to measure their endocrine and immune systems (in collaboration with immunology professor Dr. Hans Ochs), and people in the control room were coding their emotions. People adapted to the cameras and physiological recording equipment within about 45 minutes. They brought videos and music to listen to; they brought their pets; they read the newspaper; they worked, made telephone calls, prepared and ate meals, cleaned up, walked in the park, and did anything

else they wanted to. They also participated in our standard lab assessment. We followed them and repeatedly assessed them for 6 years, as 17 of them divorced, and many other couples became pregnant. We followed couples through the pregnancy, and then observed them interacting with their 3-month-old babies using a technique called the Lausanne triadic play situation, taken from Swiss psychologist Elisabeth Fivaz-Depeursigne (Fivaz-Depeursinge & Corboz-Warnery, 1999). John learned how to study babies from one of his best friends, the developmental psychologist Edward Tronick (e.g., Gianino & Tronick, 1988). Edward and John had started grad school in the same class at the University of Wisconsin in 1965. Edward spent his professional life working with America's pediatrician T. Berry Brazelton, and no one understands babies better than these two people. In John's lab at the University of Illinois and later at the University of Washington, we began with an interview we called the Oral History Interview, in which couples answered questions about the history of their relationship, their philosophy about relationships, and their parents' relationships. It turned out that couples who had many positive stories and memories to tell about their relationship and their partners' characters were the strongest; they seemed to have accurate "maps" of their partner's inner world as well (which we called love maps). The Oral History Interview was "coded" quantitatively by the "Buehlman coding system," developed in John's lab by Kim Buehlman. In another study, that coding system had 94% accuracy in predicting stability or divorce over a 4-year period (Buehlman, Gottman, & Katz, 1992).

Our 13% rate of breakup of new marriages in 6 years is pretty consistent across labs; as in our other studies, we could predict which couples would divorce and which would not (and the happiness of those who stayed married) with greater than 90% accuracy. We could predict almost perfectly how their marriages would wind up 6 years later just from their 15-minute conflict conversations with about 88% accuracy. Most of these predictions were made from the way the couples discussed a conflict issue.

Having a baby is supposed to be a blissful event. However, within 3 years after the newlyweds' first babies were born, we discovered that an astounding 67% of these couples had begun to plummet in marital happiness and increase

Natural Principles of Love

11

dramatically in hostility toward each other. If our sample is representative, for the majority of couples the arrival of the first baby is a catastrophe for their love relationship. What a huge tragedy this is! Just one-third of the couples sailed through this transition from partners to parents. What was amazing to us is that we could predict almost perfectly from data obtained a few months after the wedding whether a couple would be in the 67% group (Shapiro, Gottman, & Carrere, 2000). Couples in the 67% group also had hostility as they played with their baby, and our research found that, compared with couples in the 33% group, the baby was harmed by that hostility. These prediction levels were not small. From the way a couple in their last trimester of pregnancy talked about a conflict we could account for half the variance in how much their 3-month-old baby would laugh, smile, or cry, and the neurological ability of the baby to self-soothe, known as the baby's "vagal tone." The vagus is the tenth cranial nerve and soothes the heart and mediates the focus of attention (Porges, 2011).

When Julie and John designed a 2-day seminar based on our theory, and based on comparing the unfortunate 67% to the fortunate 33%, in a 10-hour seminar we found that we could reverse the drop in relationship satisfaction for 77% of our last-trimester couples, and we later learned that we could strengthen this effect with a support group (Shapiro & Gottman, 2004). With the Gottman Institute, we now have trained more than 1,000 workshop leaders in 24 countries, and the prevention effects replicate. Hospitals as far afield as Iceland and Australia have adopted our program.

It took us only 15 minutes of a couple's conflict discussion data to perform our couple's predictions, and furthermore, even our parameters that described how the 15-minute conflict conversation started in its first 3 minutes--a parameter we called "start-up"--predicted how the conflict discussion itself unfolded 96% of the time (Gottman, 1994). Most of the predictions we made from our initial study held across six separate replication studies, they held for heterosexual as well as same-sex couples, and they held throughout the life course. Why were these predictions so accurate? We think the reason is simple. Our lab numbers actually underestimate how negative the conflict interactions of unhappy couples are at home, and

they also underestimate how positive the interactions of happy couples are at home. We found this out by having couples in one study take the recorders home so no one else (except the camera) was present. Small effects cumulate, resulting in divergent accelerating trajectories for different groups of couples. Initial differences between the masters and the disasters are very stable. Levenson and John (Gottman & Levenson, 2002) found that there is more than 80% stability in couples' interaction over even as long a period as 14 years, even if some of the couples got therapy. Not only could we predict the fate of newly married couples over 6 years, but in Levenson's lab our prediction was even possible for couples in midlife and old age.

Let us take a moment and ask a very important question, namely, Is this divorce prediction easy? A few years ago Laurie Abraham (2013) spent an entire chapter criticizing John's methods. She claimed that if one predicted that 100% of the newlyweds in John's study would get divorced, since the U.S. national divorce rate was then 50%, John would be right half the time. So, she argued, a prediction rate of 90% accuracy wasn't that great an accomplishment. A New York Times review of her book extolled the virtues of her arguments and exclaimed that she had uncovered the charlatan who had pulled the wool over everyone's eyes (John). However, she had made a logical error. Her error was that the U.S. divorce rate has indeed been estimated by sociologists to be about 50%, but only after 40 years of marriage. These high divorce estimates have been successfully challenged by Feldhahn and Whitehead (2014). In just 6 years of marriage, only 13% of the couples in our newlywed sample divorced, so if one guessed they'd all divorce, one would be wrong by 87%. Thomas Bradbury at UCLA found a divorce rate of 7.6% over 4 years in his sample of newlyweds (Bradbury & Karney, 1993; Karney & Bradbury, 1995). So if one guessed everyone would divorce in Tom's sample, one would be wrong by 92.4%. In fact, the problem of guessing who divorces and who does not at 90% accuracy (our average accuracy across 6 separate replication longitudinal studies) in our 130 newlywed couples by chance alone is exactly like trying to pick out blindfolded and randomly 15 out of 17 red balls from a bowl that also contains 113 white balls. The probability of picking 15 out of 17 red balls correctly by chance alone can be computed as approximately 2.5 times 1015. To spell that out,

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download