Why Did People Once Sleep in Two Separate Shifts Every Night?
For many modern sleepers, waking at 2 a.m. can feel like something has gone wrong. Yet historical records contain repeated references to a different way of describing the night: people spoke of a first sleep and a second sleep, with a period of wakefulness between them. Historian A. Roger Ekirch brought this evidence together in influential research on preindustrial sleep, arguing that segmented sleep was common across much of Western history before the modern spread of artificial lighting. His evidence included diaries, court records, medical works and literature, while later historians have both expanded and challenged how those sources should be interpreted.
The basic idea is easy to picture. A household went to bed after evening activities had ended, slept for a period, woke during the night, and later returned to sleep before morning. But the popular version of the story often goes too far. It is misleading to imagine that every medieval person went to bed at exactly 9 p.m., woke at exactly midnight for two hours, and then slept another four hours. Historical sleep was more variable than that, and recent scholarship has argued that some evidence traditionally used to demonstrate a universal two-stage pattern may have other interpretations.
What the evidence does establish is more interesting than the simplified version. Nighttime was organized differently in preindustrial societies, and the darkness itself imposed practical limits on when people worked, traveled, socialized and slept. Artificial lighting gradually extended the usable day, while urbanization and changing work schedules altered when people went to bed and when they were expected to wake. The history of sleep is therefore also a history of light, labor and the changing structure of everyday life.
What “First Sleep” and “Second Sleep” Actually Meant
The phrase “first sleep” is not a modern invention created to describe an old habit after the fact. Similar expressions appear in historical texts, including references from classical antiquity. Ekirch has identified evidence for the concept stretching back to ancient literature, while research summarized on his Virginia Tech site includes Roman references to primus somnus, or “first sleep,” as well as earlier Greek descriptions of a period following an initial sleep.
In early modern English sources, expressions such as “first sleep” and “second sleep” could be used as though readers already understood what they meant. This is one reason Ekirch considered the evidence significant: the phrases often appear without lengthy explanations, suggesting that the idea of waking after an initial period of sleep was sufficiently familiar to contemporary readers. His research ultimately gathered hundreds of references from a wide range of written sources.
The waking interval did not necessarily look the same from household to household. People might remain in bed, talk, pray, think about dreams, engage in sexual activity, tend to household needs or perform other small tasks. Some accounts describe people getting up and moving around; others imply a much quieter interval. The historical record therefore suggests a period of nighttime wakefulness rather than one universal ritual.
That distinction is important because modern descriptions often turn the first sleep into a rigid schedule. The historical evidence does not justify that level of precision. People did not have alarm clocks, standardized workdays or identical household routines, and the timing of sleep depended on occupation, season, class, location and what people needed to do after dark.
Why Darkness Changed the Night
Before widespread electric lighting, darkness was a physical constraint. Candles, lamps, fireplaces and other forms of artificial illumination existed, but they were generally weaker, more expensive or less convenient than modern electric lighting. Extending the working day after sunset therefore required additional fuel, effort or expense.
That did not mean everyone went immediately to bed when the sun disappeared. People still had meals, religious observances, domestic work, social visits and other activities after dark. But the night remained fundamentally different from the modern environment in which artificial light can make midnight look almost like late afternoon indoors.
Ekirch argues that this relationship with darkness was central to the history of preindustrial sleep. As artificial lighting became more widespread, people could remain active later and postpone bedtime, gradually compressing the traditional interval between the two sleeps. The transformation was not instantaneous and was particularly uneven across class, occupation and urban environments.
The change was therefore not simply “electricity replaced two sleeps.” Gas lighting, candles, improvements in lamps, urban growth, industrial schedules and changing patterns of social life all contributed. By the nineteenth century, the organization of nighttime was changing rapidly, and sleep became increasingly fitted around a more tightly structured workday.
This helps explain why the old pattern did not disappear on one particular date. It faded through a long transition.
What People Did During the Wakeful Hours
One reason segmented sleep has become so fascinating is that historical sources describe the middle of the night as more than a period of frustration.
Some people prayed. Others reflected on dreams or spent quiet time in bed. Some carried out small domestic tasks, while others socialized or visited nearby people. Historical accounts also contain less respectable uses of the nighttime interval, including theft and clandestine behavior. The darkness created opportunities precisely because it changed who could see whom and who was expected to be awake.
Religion could make the nighttime interval especially meaningful. Medieval and early modern communities organized daily life around religious calendars, and nighttime prayer was not unusual. This should not be mistaken for evidence that religious activity caused every episode of segmented sleep. Rather, religious practice was one of the many activities that could occur when someone was awake during the night.
The mental experience of nighttime also mattered. Ekirch has emphasized historical references to dreams and the period immediately following sleep, while later scholarship has cautioned against assuming that people necessarily experienced dreams in the precise way his original model suggested. The evidence is strongest for the existence of nighttime references to waking and weaker for broad claims about a universal psychological state during that interval.
The middle of the night could also simply be quiet. That may be the most overlooked possibility. Not every waking period involved conversation, chores or dramatic reflection. A person might wake, remain still for a while, and then return to sleep. Historical language records what writers found worth mentioning, not every ordinary moment of human life.
Did Medieval People Really Sleep Like This?
Here the history becomes more complicated.
Ekirch’s research made a strong case that segmented sleep was familiar in preindustrial Western societies. References to first and second sleep appear in medieval and early modern literature, and the expressions occur across different kinds of documents. His interpretation became highly influential because it challenged the modern assumption that an uninterrupted block of sleep was always the natural human pattern.
But historians do not all agree that those references prove a single, dominant sleep schedule.
In 2023, historian Niall Boyce published a detailed reconsideration of segmented sleep in Medical History. He did not argue that divided sleep never occurred. Instead, he examined the scientific, anthropological and literary evidence behind the theory and concluded that the historical material may reflect a wider diversity of sleeping practices than the simple first-sleep/second-sleep model suggests. Some references that appear to describe two sleep periods can potentially be understood in other ways.
Boyce’s study highlights a particularly important problem: written references to “first sleep” do not necessarily document the exact physiological pattern modern readers imagine. A historical phrase can describe a period before midnight without proving that everyone later woke for an extended interval and then entered a clearly defined second sleep. In some medieval German material, for example, researchers have found references to a “first sleep” without clear evidence of a routine wakeful period separating it from another sleep.
That does not destroy the segmented-sleep theory. It makes it more precise.
The safest historical conclusion is that biphasic or segmented sleep clearly existed and was culturally recognizable in some preindustrial societies, but its frequency, timing and exact structure varied. It should not be turned into a universal law of medieval human biology.
Ancient Rome Offers a Surprising Clue
Evidence for divided nighttime sleep is not limited to medieval Europe. Classical writers used terms associated with “first sleep,” and research assembled by Ekirch includes references from Greek and Roman authors. Thucydides, for example, used the expression after first sleep, while Roman writers including Livy and others employed terminology associated with an initial period of nighttime sleep.
This is an interesting reminder that the history of sleep is older than the Industrial Revolution. Human beings did not suddenly discover the idea of waking after an initial period of sleep in medieval Europe.
The Roman evidence also fits into a broader history of how ancient societies organized nighttime activity. Urban populations could use lamps, fires and public spaces after dark, while military operations and other activities were sometimes deliberately planned around particular stages of the night. Researchers studying Roman nighttime culture have found evidence suggesting that the distinction between first sleep and later parts of the night was meaningful to Roman writers.
The broader religious landscape is another useful comparison. Ancient Romans attached strong cultural meanings to forests, sacred places and nighttime activity, as discussed in HiddenFrontiers’ article on ancient Roman fears of certain forests. The important connection is not that forests caused segmented sleep, but that ancient people often experienced night as a distinctive social and religious environment rather than simply as a dark version of daytime.
Why Modern People Tend to Assume One Long Sleep
The modern expectation of a single uninterrupted block of nighttime sleep feels natural partly because most modern institutions are built around it.
Schools begin in the morning. Offices open at specific times. Factories run shifts. Public transportation operates on schedules. Alarm clocks and smartphones organize waking time. Bedrooms are equipped with heating, cooling, blackout curtains and electric light, while supermarkets and restaurants can remain open long after sunset.
Under these conditions, consolidating sleep is convenient.
The transition was especially significant during industrialization because time became increasingly disciplined. Workers were expected to arrive at particular hours, and late-night activity was no longer merely a personal decision. Urban economies created schedules that rewarded people for staying awake at some hours and being reliably asleep at others.
Artificial lighting made those schedules possible. It allowed factories, streets, offices, shops and households to function beyond the natural limits imposed by sunset.
Ekirch’s historical work emphasizes the nineteenth century as the critical period in which sleep became increasingly consolidated, particularly in growing cities. His evidence also shows that the transition was uneven: some people retained older sleep habits long after urban residents had adopted later bedtimes and more consolidated sleep.
The result was not simply a new biological pattern. It was a new social expectation.
The Industrial Revolution Did Not Magically Create the Eight-Hour Night
Popular retellings sometimes suggest that humans once slept two four-hour periods, then the Industrial Revolution suddenly created the modern eight-hour night.
The historical evidence is much less tidy.
First, sleep duration and sleep structure are different questions. A person can sleep the same total number of hours while changing the number of waking periods during the night. Second, modern sleep science does not support the idea that all humans naturally require exactly eight hours. Sleep needs vary by age, individual biology and circumstances.
A 2015 study of three contemporary preindustrial societies in Tanzania, Namibia and Bolivia is particularly useful because it tested sleep under conditions without electric lighting. The researchers found sleep durations averaging roughly 5.7 to 7.1 hours and found that participants generally did not experience an extended waking period in the middle of the night. They usually fell asleep several hours after sunset and woke before sunrise. Temperature appeared to be an important regulator of sleep timing.
That result complicates the idea that segmented sleep is simply humanity’s original biological state.
It does not disprove the historical evidence from Europe. The societies in the modern study lived in different environments and cultures, and their sleep patterns cannot be substituted directly for historical Europeans. But the comparison shows that living without electricity does not automatically produce a first-sleep/second-sleep routine.
The strongest conclusion is therefore cultural rather than biological: limited artificial light may create conditions in which certain forms of segmented sleep become practical or common, but humans do not all respond to darkness in exactly the same way.
Climate May Have Played a Role Too
Nighttime temperature provides another possible clue.
The 2015 Current Biology study found that the three preindustrial societies studied tended to sleep during the portion of the night when environmental temperatures were falling and to wake before sunrise. Their sleep duration changed seasonally, with winter sleep lasting nearly an hour longer than summer sleep.
That finding matters because it demonstrates that human sleep is shaped by more than light. Temperature, seasonal changes, daily activity and local customs can all influence when people go to bed and wake.
Historical Europeans lived in environments where winter nights could be long and bedrooms could be extremely cold. A period of nighttime wakefulness might therefore have had a different practical meaning in winter than in summer. People still had to balance darkness, cold, household work and social expectations.
Yet climate should not be treated as the single explanation either. Ekirch’s evidence for first and second sleep appears across periods and regions with different seasonal conditions. The historical record points toward an interaction between environment and culture rather than one simple cause.
What Happened to the Two Sleeps?
The old pattern gradually weakened rather than disappearing overnight.
As urban areas expanded, artificial lighting became cheaper and more reliable. Gas illumination extended the productive evening, and electric lighting transformed it even further. People could stay awake later without depending on natural daylight.
At the same time, social schedules changed. Evening entertainment became easier to organize, commercial businesses remained open later, and industrial work imposed fixed morning obligations.
These pressures favored a later bedtime followed by one substantial sleep period.
Ekirch’s research suggests that by the nineteenth century, many people’s first sleep had begun expanding while the second sleep and intervening waking period became shorter. By the end of the century, he found examples of the traditional second sleep becoming little more than a brief return to sleep. The transformation differed significantly by social class and degree of urbanization.
In other words, modern consolidated sleep did not replace segmented sleep because someone decided that one long sleep was scientifically superior. It emerged from a changing environment in which staying awake later and waking at fixed times became increasingly useful.
The bedroom itself became integrated into an industrial society’s schedule.
Is Waking in the Middle of the Night Actually Abnormal?
The history of segmented sleep has sometimes been used to reassure people who wake during the night that nighttime wakefulness is completely normal.
There is an important truth in that idea, but it should not be turned into medical advice based solely on historical evidence.
Boyce’s 2023 reassessment specifically warns against using the historical theory as a straightforward claim that consolidated sleep is unnatural or that modern sleep problems can be solved simply by recreating preindustrial routines. The historical evidence is complicated, and literary references cannot substitute for clinical research into individual sleep disorders.
What history can show is that human beings have experienced nighttime wakefulness in many different ways. A person who wakes briefly is not necessarily experiencing something historically unusual.
But persistent insomnia, severe daytime sleepiness or other sleep problems are medical questions, not historical ones.
The useful lesson is therefore not “you should wake up every night for two hours.” It is that human sleep has never been quite as uniform as modern schedules make it appear.
The Night Was Once a Different Kind of Time
The deeper significance of segmented sleep is that it changes our picture of historical night.
Modern cities are illuminated, scheduled and continuously active. In many places, darkness no longer determines when streets become empty or when work ends. A medieval or early modern village was different. After sunset, visibility fell sharply, travel became harder, and many forms of labor became impractical.
The night became a distinct social environment.
People told stories about it, feared it, prayed in it, traveled through it, committed crimes in it and made sense of it through religious beliefs. The possibility of waking between two periods of sleep was part of that larger nighttime culture.
The phrase “first sleep” may therefore tell historians as much about how people understood time as it does about their physiology. If a society gave a recognizable name to an initial period of sleep, that suggests people were paying attention to where sleep fit within the structure of the night.
The middle hours were not simply “lost time.” They could have meaning.
What We Can Actually Say About First and Second Sleep
Three conclusions are particularly secure.
First, historical sources genuinely contain references to “first sleep,” “second sleep,” and related forms of nighttime rest. These references are found in classical, medieval and early modern material, and they formed the basis of Roger Ekirch’s influential research into segmented sleep.
Second, those references do not prove that every person followed an identical two-stage sleep schedule every night. Recent historical scholarship emphasizes ambiguity in the texts and points to a wider diversity of sleep practices than the popular version of the theory suggests.
Third, the shift toward consolidated sleep was connected to broader changes in nighttime life, especially artificial illumination, urbanization and changing work schedules. The transition was gradual and uneven rather than a sudden biological revolution.
That leaves us with a more interesting historical picture than the familiar claim that “people used to sleep differently.”
They did, but not in exactly one way.
Why the Story Still Matters
The appeal of first sleep and second sleep is partly nostalgic. It offers an image of a world in which people went to bed early, woke during the night and spent a quiet hour thinking, praying or talking before returning to sleep.
Some of that picture is supported by historical evidence. Some of it is interpretation. Some has been amplified by modern writers who like the idea that our ancestors possessed a more “natural” relationship with sleep.
The evidence does not support such a simple romantic conclusion.
What it does show is that sleep is deeply shaped by society. People do not sleep in a vacuum. Lighting, work, climate, housing, religion, family life and expectations about time all influence when the day ends and the night begins.
That is why the history of segmented sleep is ultimately less about proving that everyone once woke at midnight and more about recovering a lost way of thinking about nighttime. The modern expectation of continuous sleep feels timeless because our institutions reinforce it every day. Historical records reveal that human beings once organized the hours of darkness differently—and that the way we sleep can change when the world around us changes.
The next time someone wakes in the middle of the night, there is no need to imagine that they have been transported back to medieval Europe. But there is also no reason to assume that a brief period of wakefulness is somehow alien to the human past. Our ancestors left behind evidence that the night could contain more than a single uninterrupted block of darkness and dreams.
Their first sleep ended. The night continued. And, eventually, they went back to sleep.
SOURCES & REFERENCES
Ekirch, A. Roger — “Sleep We Have Lost: Pre-industrial Slumber in the British Isles,” American Historical Review, 2001.
Foundational historical research behind the modern academic discussion of first sleep, second sleep and preindustrial segmented sleep.
Read the Cambridge discussion and citation of Ekirch’s study
Ekirch, A. Roger — “Segmented Sleep in Preindustrial Societies,” Sleep, 2016.
Ekirch’s response to anthropological sleep research and further discussion of historical evidence for segmented sleep.
Read the article on PMC
Boyce, Niall — “Have we lost sleep? A reconsideration of segmented sleep in early modern England,” Medical History, 2023.
Important scholarly reassessment emphasizing uncertainty, alternative interpretations and diversity in early modern sleep practices.
Read the full Cambridge article
Yetish, Gandhi, et al. — “Natural Sleep and Its Seasonal Variations in Three Pre-Industrial Societies,” Current Biology, 2015.
Scientific study of sleep patterns among three contemporary preindustrial societies in Tanzania, Namibia and Bolivia.
Read the study
Dijk, Derk-Jan & Skeldon, Anne C. — “Human sleep before the industrial era,” Nature, 2015.
Scientific commentary on what the preindustrial sleep research suggests about sleep duration and environmental regulation.
Read the Nature article
Roger Ekirch, Virginia Tech — Segmented Sleep research archive.
Researcher’s collection of historical references and supporting material concerning first and second sleep across periods and regions.
Visit the research archive
Scientific American — “Ancient Sleep in Modern Times.”
Useful historical-scientific overview of Ekirch’s original argument and the relationship between artificial lighting and nighttime sleep routines.
Read the Scientific American article
FAQ
Did people really sleep in two shifts in the past?
Historical sources genuinely contain references to “first sleep” and “second sleep,” and historian Roger Ekirch has documented hundreds of such references. However, recent scholarship cautions that the evidence does not prove everyone followed an identical two-stage schedule every night.
What was first sleep?
“First sleep” generally refers to an initial period of nighttime sleep before a later period of wakefulness and, in the historical model, a subsequent “second sleep.” The exact timing and duration varied rather than following a universal timetable.
What did people do between first and second sleep?
Historical references describe a wide range of activities, including prayer, conversation, reflection, household tasks, sexual activity and other forms of nighttime behavior. Some people may also have simply remained quietly in bed before returning to sleep.
Did medieval people go to sleep immediately after sunset?
Not necessarily. Although artificial lighting was more limited than today, people still worked, ate, socialized and traveled after dark. Ekirch’s evidence indicates that bedtime could vary considerably according to season, occupation and what people had to do.
Did electricity end segmented sleep?
Artificial lighting was an important factor, but it was not the only one. Gas lighting, urbanization, industrial work schedules and changing social habits also helped shift sleep toward later bedtimes and more consolidated nighttime rest during the nineteenth century.
Is segmented sleep more natural or healthier?
The historical evidence does not justify saying that segmented sleep is inherently healthier. A 2023 scholarly reassessment specifically cautioned against turning the historical theory into medical advice, while modern research shows that sleep patterns vary across populations and environments.
Did people in ancient Rome also have first and second sleep?
There are references to “first sleep” in Greek and Roman sources, including terminology used by writers such as Thucydides and Roman authors. This suggests that the concept of an initial period of nighttime sleep predates medieval Europe.
Why does the idea of two sleeps remain so popular?
It challenges the assumption that today’s sleep schedule is timeless and shows how strongly sleep is shaped by technology, work and social expectations. The appeal is real, but the historical evidence is more varied than the popular two-four-hour/two-four-hour version of the story.
