Only in the past four decades have we seen gender norms shift in a radical way. Until at least the 1960s, Americans defaulted to convention: Men made money; women made babies and homes. The gender line, psychologists learned, even transcended waking life. Men and women not only faced different social and economic expectations; they dreamt differently, too.
In 1966, psychologists Calvin S. Hall and Robert Van de Castle analyzed thousands of written dream reports (submitted by dreamers themselves) and identified certain types of characters, scenarios and emotions as distinguishing elements of male and female dreams. Compared to men, women recounted dreams with more characters and positive social interactions. And men were more likely than their xx-chromosome counterparts to report situations and feelings associated with aggression and achievement.
These male-female generalizations consistently bore out in dream studies — until they didn't. Researchers started to find that dreams grew less reliably gendered as societal gender roles blurred. In one recent study, psychologists from the University of Ottawa sought to see if years-old results about male and female dreams held up, by having human beings, as well as a computer program, predict study participants' genders based solely on their dream reports.
Sigmund Freud and his psychoanalytic pals saw dreams as manifestations of suppressed urges and conflicts. To Freud, every character, object and event in a dream held symbolic meaning. But, by the second half of the 20th century, plenty of psychologists challenged Freud's theory of dreams. The continuity hypothesis, introduced in the 1950s by Calvin Hall, says dream content (what we dream about) is mainly an extension of our waking thoughts and concerns. The average dream of the average dreamer, in Hall's view, would be heavy on minutiae.
To study dream content in a quantifiable way, Hall and colleagues devised a dream-coding system. It worked like this: Volunteers wrote summaries of their dreams. Then, researchers reviewed their "dream reports" for elements including characters (further categorized as humans, animals or imaginary creatures), emotions (e.g., happiness, anger, confusion), activities, types of social interactions (e.g., friendly, aggressive, sexual), physical objects, settings and scenarios marked by striving or fortune. Each element corresponded to a specific code.
Through dream-content analysis, researchers have pinpointed defining dream elements for specific groups of people, including young children, end-of-life patients and, of course, men and women. Dream reports from the '60s through the early '80s offered a fairly consistent picture of male and female dreams. Male dream reports tended to mention physical aggression, the outdoors, unfamiliar settings, tools and weapons. By comparison, women's dream reports contained more characters and references to social aggression (trash-talking), emotions, clothing and household objects.
Then, in 1985, researchers compared the dreams of working mothers and homemakers, to see if the expected "female" themes would emerge, whether or not participants did traditional lady-things. The verdict? Nope. "As the trend toward carrying the dual role of wage-earner and homemaker is expanding," study authors wrote, "the gender differences typically observed in dreams content may decrease." In other words, women who filled "men's roles" were less likely to dream like girls.
Since then, dream-content analysis has continued to show that social realities seep into the #content churned out by the subconscious (aka, dreams). In the current study, researchers yet again assessed whether or not gendered dream elements identified in years-old studies still hold up. But, in addition to coding dream reports themselves, researchers asked volunteers to read the reports and predict the gender of the study participants who wrote them. They also pitted humans against a computer program built specifically to score dream reports. The "machine learning algorithm" scanned reports for language associated with "male" and "female" themes and emotions. Other research groups have used computer programs to analyze dream content, but not with the goal of predicting gender.
For the study, 100 participants, all between 18 and 24 years old, were supposed to wake up and write down their nightly dreams over a 10-day span, making sure to include details about locations, events, characters, feelings, activities and social interactions. Four human judges, who fell into the same age range as participants, read each report (288 in total) and determined whether they thought a man or woman had written it.
The judges had no formal background in sleep, dreams or gender dynamics, and they didn't try to code the reports in any systematic way. But many of the dream reports contained obvious gender clues, such as a dreamer calling himself an Indian man. So, researchers omitted the clues, and asked a different set of judges to read, and make gender predictions about, the revised, neutered reports. (It's worth noting that, as in a lot of science research, "sex" and "gender" aren't clearly delineated.)
The human judges beat the computer program, but not by much: Humans predicted gender correctly almost 75 percent of the time, on average, compared to the program's accuracy rate hovering just below 70 percent. But, human accuracy suffered when researchers removed clues from the reports, whereas the computer program performed similarly with or without the tip-offs.
What gave the human judges a leg up? Familiarity with stereotypes. "[Human] judges were able to look at the overall tone of the entire dream report," study authors wrote. "Despite this, the judges acknowledged the ambiguity of the majority of dream reports and admitted to making their decisions based on traditional gender stereotypes," e.g., a dreamer mentioning a boyfriend.
Going into the study, researchers specified a number of elements they assumed would distinguish dreams as male or female, but only one of their guesses panned out: Male dream reports contained more references to quantity. "The gradual decline of gender distinctions in waking-life societal roles over more recent years may have influenced this corresponding change in dream content," researchers wrote.
The human-versus-machine aspect of the study is worth noting, but it shouldn't outshine the other takeaway: Dreams, like sleeping habits, tell us something about the world we live in and the social norms we live by.
Sleep researchers, for example, have suggested that women disproportionately suffer from insomnia partly because they're conditioned to be caretakers, and their instinct to tend to others' needs keeps them up at night. But in this case, decreasingly predictable differences between the stuff of male and female dreams reminds us that we have made progress over the last 40, or even 15, years in razing long-standing gender norms. Now, all we need to do is make sure we protect people whose lives, and dreams, don't fit the mold.