The Five Love Languages
Tell someone that our society has been profoundly transformed by statistics and they may nod politely, perhaps stifle a yawn.
Then listen to what they want to talk about.
You might find yourself discussing artificial intelligence, the latest Biden-Trump poll, Ozempic, or the chances of rain next week: Topics that wouldn't exist without statistics.
We're all familiar with statistical data – poll results, student test scores, inflation trends, Caitlin Clark's field goal percentages, the probability of this, the rising tide of that.
Less noticeable, for most people, is the role of statistics in science and technology, from drug testing to global warming estimates to the development of large language models that fuel generative AI.
Least noticeable of all are the influences on the way we think. For instance, consider this factoid from the U.S. Census Bureau:
Marriages that end in divorce last about eight years on average.
For us, there's nothing particularly hard to understand about this statement, because we routinely discuss averages. Prior to the late 19th century, statements like this would've baffled readers. Mathematical averages were used, on rare occasions, by scientists, but almost never by anyone else.
It's not that people knew what a mathematical average is but lacked a term for it, such as "average" or "mean". Rather, averages just weren't calculated very often, much less pondered. (Aristotle's "golden mean"? It's more like a median – a desirable mid-point between extremes.)
Why am I geeking out about this? Because this newsletter illustrates that even pop psychology has been statisfied, or transformed by statistics, over the past century.
Specifically, I'll be describing a new review of studies on the Five Love Languages (5LL). Although relationship advice has been around for millennia, the 5LL theory couldn't have been articulated until the 20th century, because it's grounded in statistical concepts and procedures.
Why is this important?
If there's a "secret" to a great relationship, it's clear that some people (including me) either don't know it or aren't able to put it into practice. Although divorce rates in the U.S. have been declining in recent decades, nearly 40% of marriages still end in divorce. Other evidence, including informal conversations with friends (and, in some cases, personal experience) remind us that not everyone who's partnered is satisfied with the partnership. Some scholars, most notably Bella DePaulo, argue that single people are generally better off.
My hope is that this newsletter contains some useful advice for present and future couples – even through the 5LL doesn't hold up well under scientific scrutiny, and there probably isn't some "secret" to what you'd consider a satisfying relationship.
The 5 Love Languages
The 5 Love Languages, a perennial bestseller following its initial publication in 1992, has now sold more than 20 million copies worldwide. The author is Dr. Gary Chapman, a Baptist pastor with a background in anthropology and adult education.
In essence, the love languages are five ways of giving and receiving love:
words of affirmation (praise, "I love you", and other expressions of affection)
gifts (particularly those that are thoughtfully chosen)
acts of service (doing the dishes, helping with child care, running errands, etc.)
quality time (being fully present while interacting)
physical touch (holding hands, hugging, being sexual, etc.)
If Chapman had merely called these five ways to express love, nobody would disagree with him – or pay much attention either. Successful theories don't restate the obvious. What distinguishes the 5LL theory is three additional assumptions:
Assumption 1: Each person has one primary love language.
Assumption 2: There are exactly five love languages.
Assumption 3: People whose love languages match have better relationships.
From these assumptions, it follows that the "secret" to an enduring, happy relationship is for each person to better understand and speak their partner's love language.
These assumptions are the foundation of a hugely popular commercial enterprise. On the 5LL website you can find love language quizzes, special events, instructional resources, and, inevitably, a store (consisting mainly of books). Almost everyone has heard of the 5LLs, and Dr. Chapman maintains a nationally-syndicated radio show as well as the ultimate hallmark of fame: an appearance on Oprah. (Her love language was said to be words of affirmation.)
Statistics plays two key roles in the 5LL framework.
(a) Love languages are determined by simple tests that rely on 20th century statistical methods. On the 5LL website there's a free version of the love language test as well as a paid premium assessment ($35) which promises a better relationship in 20 minutes, with "no hard work".
(b) Relationship satisfaction is assumed to be both quantifiable and predictable from another variable (whether or not love languages match). As a result, Assumption 3 (People whose love languages match have better relationships) is easy for us to understand, but would not be very comprehensible to our pre-statistical predecessors.
So, are the three assumptions justified? Is the 5LL a sensible theory? Can it actually help improve anyone's relationship? (As you'll see, these are somewhat independent questions.)
A new review
For the April 2024 issue of Current Directions in Psychological Science, Dr. Emily Impett at University of Toronto and colleagues reviewed 5LL studies pertaining to the three key assumptions.
Assumption 1: Each person has one primary love language.
This assumption doesn't pan out.
As Impett and colleagues put it, "studies have consistently demonstrated that people tend to endorse all five love languages as meaningful ways of expressing love and feeling loved."
Specifically, when adults are asked to rate on a scale from 1 to 5 how desirable or meaningful they consider each love language, the average ratings come in around 4. Most people consider each love language to be important.
Crudely speaking, the problem here is that the measure is rigged. Chapman's love language test forces respondents to choose between languages. Here's a typical item:
It's more meaningful to me when...
I get to do things with someone I love.
[versus]
I hear supportive things from someone I love.
Questions like this drive me nuts. I'm not sure what "meaningful" means here exactly, but let's suppose the question is about what I prefer, or what I consider a deeper or more desirable expression of love. My answer is: It depends. Sometimes I appreciate doing things with my partner, sometimes I appreciate hearing supportive things. (Often it's nice to have them at the same time.)
In prior studies, the primary love language revealed by Chapman's test was not consistently associated with attitudes towards each love language, when people were allowed to rate each language separately instead of being forced to choose between them.
For instance, on Chapman's test, gifting turns out to be the primary love language of fewer than 5% of people, but gifting tends to be rated fairly highly as an expression of love. Here's how Impett and colleagues explain this:
"[I]f they had to choose, most people might favor spending quality time with their partner over receiving gifts, as it is something they can do more frequently with their partner in daily life, but if they were to independently rate each love language, they might rate receiving gifts as quite meaningful because it represents an occasional but special way in which their partner communicates their love."
The problem with the 5LL test comes up in other pop psychology frameworks, such as the Myers-Briggs typologies. Some of the M-B personality tests force you to choose between options (Are you more influenced by peoples' stories or by factual data?), but the answer you probably most want to give is: It depends.
In sum, the first assumption of the 5LL is a sort of a caricature. Not that many people seem to have a primary love language. It only seems that way when we're forced to make choices that oversimplify how we feel.
Assumption 2: There are exactly five love languages.
Impett and colleagues find the data on this assumption to be inconsistent and unreliable. It's unclear how many love languages there are or how they might be described.
In my opinion, it's not useful to try to nail down a set of discrete "love languages". This feels to me very much like a game of finding sufficiently inclusive terms.
For instance, I would say that there are exactly five love languages: Desire, affection, kindness, gratitude, and commitment.
I truly believe this. I believe that every possible expression of love could be stuffed into one of those five categories. But I don't think this is an especially insightful taxonomy. I've just chosen labels that are broad enough to encompass every expression of love I can imagine.
Assumption 3: People whose love languages match have better relationships.
This assumption doesn't pan out either. For instance, partners whose love languages match (as measured by Chapman's test) don't report greater relationship satisfaction, on average.
Impett and colleagues find this unsurprising, in light of data showing that relationship satisfaction isn't predicted by matching on other dimensions, such as verbal style. This is not to say that opposites attract, but merely that similarities often fail to predict relationship satisfaction.
In the end, I find the 5LL approach potentially harmful as well as helpful.
The negatives
Judging from articles in the national news, you might consider the 5LL an innocuous bit of fluff, in the sense that if it's too superficial to help anyone, it certainly won't hurt. I disagree. Here are three potential sources of harm:
1. Lack of inclusivity.
5LL theory is pitched to heterosexual married couples. There's nothing wrong with that per se – such couples do need support – but it implies that if you're not straight and married (and, in many of the books, Christian) you're somehow not doing it right, or not worthy of the message.
Polyamory is getting a lot of attention in the news lately as a minor but growing trend. Practitioners would also find themselves excluded from the 5LL framework. (If I were a polyamorist, my love language would be logistics. I'm only half-joking.)
2. Superficiality.
Linda Carroll, a marriage and family therapist, remarks that some couples she works with treat the 5LL theory as a quick fix, expecting it to heal their relationship issues with minimal time and effort. (The 5LL website and at least some of the books do promote this expectation.) Impett and colleagues note that
"strictly adhering to the love-language label may lead [couples] to undervalue other expressions of love outside of their “primary” love language, dismiss the full range of emotional needs and preferences that go beyond the limited scope of five love languages, or discount potential or current partners who do not “match” their primary love language."
The 5LL website promises you can "learn to identify the root of your conflicts" but that's only true if the problem is simply that, say, your partner wants physical intimacy all the time and never wants to do the chores (or vice versa), but you don't feel the same way. Not all relationship problems boil down to the way love is expressed.
3. Failure to consider negative languages.
The 5LL approach skips over love languages that might be universally toxic. For instance, Dr. John Gottman is well-known for longitudinal data showing that the most accurate predictor of whether couples (both straight and gay) will get divorced or not is contempt toward the other person. 5LL tends to overlook the fact that some languages are best left unspoken.
In sum, in the worst-case scenario, 5LL creates unrealistic expectations, distracts couples from the relationship work they need to do, and downplays potentially harmful languages.
5LL theory seems to be most useful when it's treated as part of a broader effort to strengthen a relationship rather than the sole method of doing so.
The positives
1. Appreciation of one's partner.
Experts, including Impett and colleagues, often note that the 5LL fosters communication, helping people recognize that their partner may have different needs from their own. Anything that increases appreciation of who your partner is and what they need seems like a good start toward improving a relationship.
2. Encouragement.
I spoke or texted with several professionals who do couples counseling. They all report good impressions of the 5LL theory, although none of them proactively use it. Instead, when clients mention the theory, these professionals find that it stimulates helpful conversations. As Carly Wielstein, a marriage and family therapist, told me, couples may "just like the idea of there being an answer or reason for their communication breakdown." In other words, the simplicity of the 5LL approach may encourage couples to continue working on the relationship.
3. The healthy-diet metaphor.
Although 5LL assumptions aren't supported by scientific evidence, Impett and colleagues did glean a positive message from the data. They call it the healthy-diet metaphor.
Just as we need a mix of nutrients to achieve optimal health, so we need a mix of love languages for the best possible relationship. And, people have different needs at different times. Sometimes your partner needs to fix the showerhead instead of telling you, again, how much they love you. Sometimes they need to stop fooling with the showerhead and shower you with affection instead.
Gideon Park, one of the co-authors of the new review, remarked to me via email that "while this metaphorical framework emphasizes the importance of addressing various relationship needs, factors like contempt are indeed influential risk factors and can be detrimental to relationship health."
In other words, just as a healthy diet depends on avoiding avoid toxic substances, so a healthy relationship not only depends on a mix of love languages but also on avoiding toxic interaction styles and behaviors. More of the good, less of the bad.
Impett and colleagues don't present their metaphor as the "secret" to better relationships, but simply as a more helpful way of thinking about love languages. As they note, much more is needed than adjusting your particular language.
Final thoughts
Statistics can help evaluate popular relationship advice like the 5LL, and by telling us about relationship dynamics among large groups of people. For instance, a study published this March showed that feeling known by one's partner is a stronger predictor of relationship satisfaction than feeling that one knows one's partner. That resonates. On other hand, John Gottman and colleagues have found that the "Four Horsemen of the Relationship Apocalypse" are contempt, criticism, defensiveness, and stonewalling. That resonates too, painfully.
These are generalizations from group data. As Gideon Park told me, "It’s crucial to appreciate the general patterns of research findings...while also acknowledging the unique dynamics of individual relationships that researchers continue to explore." You and your partner still have to find ways to meet each others' particular needs.
In the end, the 5LL approach makes this task sound too easy. Go to the 5LL website and the first thing you see on the home page, in enormous purple font, is the declaration that "Relationships don't have to be complicated."
What do you think, dear reader? If you and your partner's personalities and lifestyles mesh seamlessly, neither one of you ever loses your temper, you don't have financial pressures or health issues or fussy toddlers or meddling in-laws, and both of you want to watch the 3 Body Problem at exactly the same time, then sure, maybe your relationship doesn't have to be complicated.
For the rest of us, relationships are complicated. (Every human being is complicated; put two of us together and how could the complexities not increase?) Although the 5LL theory can stimulate helpful discussions with a partner, we can't reduce relationships to formulas or improve them in a formulaic way. Maybe the fact that we can't do that is something to bond over. Maybe a relationship is more readily strengthened if you're both assuring each other: This is hard!
Thanks for reading!