Understanding the debate
over segmented sleep

For eight uninterrupted hours each night? In two chunks, separated by a late-night break for work, chit-chat and procreative fun? However we effing please, provided we log between six and nine hours in a given 24-hour period?

We know that human sleep has changed over the past few millennia, but we don't know exactly how the changes relate to cognition or social and technological progress or — here's a big one — what the changes say about humans' natural mode of rest.

If we discount fringe positions, two main schools of thought control the conversation. In one corner, we have experts arguing that our current concept of "normal" rest — one stretch of sleep, starting in the pm — is a glitchy byproduct of modernity; a wrong turn off evolutionary road. The other side kindly asks that we stop fetishizing "ancestral sleep," contending that it's not clear we'd rediscover biphasic sleep (i.e., two sleep sessions) if we ditched bright lights and big tech.

Researchers use various types of evidence to support their positions. Evolutionary scientists look at differences between sleep in humans and other primates, namely the six other species classified as great apes (aka hominids). Other researchers have tackled the issue by analyzing sleep habits in contemporary hunter-gatherer groups. And then other experts draw heavily on historical records of human sleep, seeking to understand how social and technological advancements have influenced shifts in resting habits. Unsurprisingly, different approaches yield different answers. While both sides boast strong evidence and credible proponents, no one has an airtight case. In a recent review paper, a team of evolutionary anthropologists and psychiatrists at Duke University summarized this debate, among other lingering what-ifs in evolutionary sleep.

About two million years ago homo erectus grew too large for his tree-branch bed and started sleeping on the ground, a shift considered a turning point for cognitive and social development. "Freed from the disadvantages of arboreal sleep they could have achieved longer duration and higher quality sleep, which would have improved waking cognition," wrote Charles Nunn, David Samson and Andrew Krystal, who co-authored the Duke paper.

Humans share certain sleep-related behaviors with other great apes, all united by our most recent common ancestor — a term of art in evolutionary anthropology — who lived about 14 million years ago. For example, while we're the only species to doze on pillow-topped mattresses, we're hardly the only species to build elevated sleep environments. All great apes make platforms or nests to sleep on, a practice that probably emerged between 18 and 14 million years ago.

But, in some ways, human sleep is unique: All humans sleep on the ground. In other great-ape species, large males (and large males alone) only hit the floor when risk of predation is low.

Humans also stand out in terms of sleep architecture, a term for the structure and arrangement of sleep stages. Compared to other primate spacies, humans spend the least amount of time sleeping (seven-or-so hours, on average) and spend the greatest proportion of their rest in REM sleep, the phase when most vivid dreaming occurs.

While sleep has changed across all great apes, human sleep has evolved disproportionately. We know this because scientists can use statistical modeling to determine whether or not one species in a family (e.g., human beings in the great-ape family) has evolved in a comparable manner to other species in the family. In other words, could we look at changes in the sleep (schedule, environment, architecture, etc.) of chimps & company over the past few millennia and accurately predict the course of human sleep? Nope. When it comes to sleep, humans are outliers. Evolution-based predictions of how humans should sleep in 2016 differ considerably from how we actually sleep.

So, we know that human sleep has evolved — extensively — in concert with changes in physical size, social structure, living environments and other behaviors and characteristics. Now we're getting to the big questions: Was human sleep supposed to change so much? Can we point to a time in history when humans slept the way we "should?" What's the natural way for homo-homos to hit the sack-sack? Is there one?

Here's one theory, introduced into mainstream thought in 2005 by the historian E. Roger Ekirch: Biphasic (or segmented) sleep is "natural." Artificial light killed humans' natural habit of sleeping in twice-daily segments (just like video killed the radio star). Let's backtrack. At some point, in some civilizations, our ancestors maintained segmented (or biphasic) sleep schedules. They turned in for their "first sleep" around 9pm or so, woke up in the middle of the night, hung out for a few hours, and then went back to bed for a "second sleep." Biphasic sleep got phased out during the industrial revolution, when artificial light became cheap and easily accessible.

But, while the demise of biphasic sleep (as a norm) is well-documented, its origins are less clear. It might be hard to refute that segmented sleep was all the rage leading up to the advent of machine tools. But what about before then? Do we have a good reason to believe that early humans slept in shifts? Evolutionary scientists aren't so sure.

Here's another theory, favored by the evolutionary crew and based on contemporary comparisons of "primitive" and "modernized" societies: With or without electrical outlets or augmented reality, humans would sleep a lot like most people do today — in one stretch, every night, for about seven hours. Evidence for this view comes courtesy of a UCLA study published last year. A research team, lead by psychiatrist and sleep researcher Jerome Siegel, compared the sleep habits of three modern-day hunter-gatherer peoples who, despite being around in 2016, maintain a tech-free lifestyle similar to that of pre-industrial humans. Data from actigraphy trackers (i.e. fitbits), the study found, suggested that all three groups slept a lot like today's i-phone toting humans: in one, nightly seven-to-eight hour block, with little daytime napping. The study also showed that bedtimes varied considerably more than wake-times did.

If we accept this theory, what can we make of the pre-industrial biphasic craze? Segmented sleep may have arisen when humans migrated away from the equator and experienced longer periods of darkness, as Siegel told The Atlantic last year: “The long nights caused this pathological sleep pattern and the advent of electric lights and heating restored the primal one.”