Socially Thermoregulated Thinking: Temperature Fluctuations Drive Our Need For Loved Ones

Harry Harlow. Most of us know him for one of the more unethical contributions a psychologist has made: “The Nature of Love.” In it, he describes rhesus monkeys raised by a terry cloth or wire surrogate mother. The rhesus monkeys preferred the terry cloth mother over the wire mother, even when the latter provided food. Further, those “raised” by the terry cloth mother were more exploratory and playful. From these findings, Harlow concluded that soft touch is essential for optimal functioning.

In a much lesser-known (but not less unethical) contribution, “The Nature of Love – Simplified,” he described findings concerning a rhesus monkey being raised with a surrogate mother that either radiated or that was devoid of any physical warmth. The monkey that was raised by a mother devoid of any physical warmth (compared to the one radiating warmth) yearned for much less contact with a neutral surrogate mother in the face of threat two months later.

However unethical these studies may have been, they suggest a fundamental fact of many animals’ existence: the desire and need for literal physical warmth and avoidance of caregivers if warmth was missing in early development. Our own studies on human beings have revealed that being confronted with a threat (being physically cold) also makes us yearn for human contact, but the desire for human contact only arises if we are familiar with safe and secure relationships.

Researchers that study temperature regulation and social contact are part of a domain called social thermoregulation. The “simple” definition of social thermoregulation for humans is that modern social relationships are plesiomorphically organized around processes of body temperature regulation. More intuitively stated: our modern social relationships are rooted in our “penguin bodies.” Like penguins, humans are homeothermic endotherms (warm-blooded animals that generate heat internally).

But regulating temperature internally is costly and generally pretty inefficient. Nature has equipped – through years of evolution – a much more efficient mechanism: homeothermic endotherms (like humans, dogs, chimps, and cats) outsource temperature regulation to conspecifics. Because we can outsource, we share the risks and load of having to deal with colder environments. In addition, we tend to lose heat much less quickly (after all, if we huddle up, our volume becomes bigger relative to our surface, from which we lose heat).
In humans, this general area of research first emerged in 2008, when Lawrence Williams and John Bargh handed participants warm versus cold cups to participants and these participants judged others as more sociable after having held a warm cup, while they themselves became more generous. In the past few years, however, psychology has been working on cleaning up its act, and many findings have not held up to scrutiny. Studies in the field of social thermoregulation have been a mixed bag: Some findings have not held up to scrutiny (see e.g., here and here), whereas others have (see e.g., here and here). One of the often-heard (and justified) criticisms of psychology studies is that the sample sizes are too small (which included my own work). This is not always true for social thermoregulation studies (see e.g., here and here). Furthermore, it has often been presumed, but not reliably or empirically demonstrated, that social thermoregulation in humans is related to their interpersonal attachments, and the relative feelings of safety in relationships (like the rhesus monkey that was raised with a warm wire mother).

We thus set forth to conduct studies on social thermoregulation that were 1) sufficiently powered to detect effects, 2) better executed than we did before, and 3) better documented, so that colleagues could scrutinize our data, while we 4) explored social thermoregulation’s relationship with people’s feelings of security in relationships. Here’s what we did. We had two experimenters approach participants in the university cafeteria at Tilburg University in the Netherlands. One of them handed participants a warm or cold cup. This experimenter did not further interact with the participant. The other experimenter (blind to which cup the participant had received), supported by instructions on the computer screen, interacted with the participant. Before having held the warm or cold cup, the participant was asked 36 questions about their interpersonal attachments. After having held the cup, the participant was asked to recall five people as quickly as possible. Following this recall, the participant was asked how close he or she felt to the recalled individuals.

A priori (and based on findings by Williams and Bargh from 2008), we had predicted that if participants held a warm cup that they would have recalled people psychologically closer to them. Turns out we were mostly wrong. We found the opposite: when participants had held a cold cup, they reminded themselves more of people close to them. Because this finding was unexpected, we were determined to replicate this finding. In a second study, we found the same effect (total number of participants for Studies 1 and 2 was 366).

But part of the story is still missing. Indeed, a desire for loved ones when cold should hinge on people’s experiences with security in relationships in the past. Unfortunately, in past studies, the link between temperature and feelings of security in past relationships is not as simple as we had hoped it to be. In one of our (underpowered) studies, we found a relationship and in another study we could not find a relationship. We suspected that part of the explanation could be that the average (pretty abstract) question about safety in relationships was pretty far removed from (the very concrete desire for) social thermoregulation (e.g., “I tell my partner just about everything”).

One aspect that psychologists don’t often do is looking at their data and generate hypotheses from them. This is exactly what we did in our project. We look at each of those individual 36 questions’ relationship with temperature and the outcome, the extent to which we think of loved ones. It turned out that only three of the questions were relevant to social thermoregulation (namely, “My romantic partner makes me doubt myself”; “I feel comfortable sharing my private thoughts and feelings with my partner”; “I find it easy to depend on romantic partners”). We generated these from our Study 1. Scientifically, this now became a hypothesis. We thus had to replicate the finding in our Study 2 and we indeed detected the same effect.

At that time, we had submitted these studies to the Journal of Experimental Social Psychology. The editor and reviewers at that time were skeptical of the study and its replication (perhaps because of the scrutiny that this subfield has seen). They thus requested us to do another study. Following this review process, we pre-registered (i.e., deposit our hypothesis and analysis script) on a website, the Open Science Framework. Because I had moved to France in the meantime, we conducted the next study in France. The pattern we detected was again largely similar: Participants in a cold condition tend to think more of loved ones, and this effect occurs for those with past positive relationship experiences. The data and analyses are open for all who would like to scrutinize our findings or who would like to reuse our data.

All in all, we collected three well-powered studies and we found – contrary to our initial expectations – that holding a cold cup in one’s hand leads someone to think of loved ones, and this really happens if one – like Harlow’s rhesus monkeys – has had relationship experiences that were warm. These findings are now part of a larger collection of findings that show, for example, that people who live further from the equator protect their core body temperature by having more diverse social contacts. We thus concluded that we form attachments in part because – early in life – our caregivers were able to provide physical warmth when we so desired, just like Harlow’s monkeys. And we further suspect that modern social relationships still very much rely on temperature regulation. Many of the mechanisms we do not fully understand; modern technology at least now finally allows us to investigate more complicated mechanisms. It is also very likely that we will take a few more wrong turns along the way of studying, but we are becoming more and more convinced that thermoregulation is an important factor in shaping human relationships.