Sleep is perhaps the most mysterious part of the human experience, and also one of the most ubiquitous. For what's apparently meant to be 8 hours a day, it seems like most people turn off our minds, relax, and float downstream. We lose consciousness of our surroundings, and intermittently experience odd and sometimes very vivid hallucinations called dreams. This is a weird thing to do, but it's only relatively recently that people have been researching the physiology, psychology, and history of sleep. As a result, it's difficult to tell whether what we do now is natural or has been influenced by electric lights and de rigueur city living. Maybe sleep was different before modern life?

The historian of sleep who is prominently pushing the hypothesis that modern sleep patterns are due to the industrial revolution is Roger Ekirch. He wrote a 2001 article called Sleep We Have Lost: Pre-Industrial Slumber In The British Isles (the journal seems to have made this article freely available, which is nice of them) and extended his thesis geographically and temporally into a 2005 book, At Day's Close: Night In Time's Past. And to cut a long story short, Ekirch finds a wide range of references in historical texts to what he calls 'biphasic sleep' or 'segmented sleep'.

Ekirch's argument is that in the past, people experienced 'biphasic sleep': according to Ekirch, numerous bits of pre-Industrial revolution writing discuss 'first sleep' and 'second sleep'. 'First sleep' would start reasonably soon after nightfall, and would continue until around the middle of the night - midnight. Then, people would wake up, and would do things (rather than just lie there trying to get back to sleep) - socialising, sex, talking, etc. After this waking period, there would be 'second sleep', which would last into the morning.

To give an example which features a post-Industrial Revolution person wondering at the practice of people living outside of cities and gaslight or electricity, Ekirch in his 2001 article quotes from an 1878 journal of Robert Louis Stevenson (the author of Treasure Island). Stevenson is the product of an industrial society - he was born in Edinburgh, Scotland - and wrote this journal while journeying through the not-very-industrial French countryside. He marvels at the sleep habits of the people around him:

"There is one stirring hour," he later recorded in his journal, "unknown to those who dwell in houses, when a wakeful infiluence goes abroad over the sleeping hemisphere, and all the outdoor world are on their feet."

Stevenson also points out that this mid-night phenomenon is so taken for granted in the French countryside that nobody even bothers to try and explain it:

"At what inaudible summons are all these sleepers thus recalled in the same hour to life?....Even shepherds and old country-folk, who are the deepest read in these arcana," he marveled, "have not a guess as to the means or purpose of this nightly resurrection. Towards two in the morning they deciare the thing takes place; and neither know or inquire further."

The philosopher Isaiah Berlin famously argued that some scholars were 'hedgehogs' and some scholars were 'foxes'; for Berlin, 'hedgehogs' believe that there is one big thing that explains everything, while 'foxes' tend to draw on a wider range of experiences and resist boiling everything down to some ways. In some ways, Ekirch seems like a classic hedgehog - it seems like he sees the biphasic sleep everywhere. This might well be because it actually was everywhere. It also might be that everything he sees just looks like biphasic sleep, but he's taking things out of context or interpreting them in ways favourable to his argument.

In a 2015 paper in Current Biology, Jerome Siegel and colleagues went to test Ekirch's pre-Industrial biphasic sleep hypothesis by examining the sleep cycles of 'pre-industrial societies' in Tanzania, Namibia, and Bolivia. I'm a bit dubious about calling people whose behaviour was investigated for a 2015 paper 'pre-Industrial' - 2015 is definitely post-Industrial revolution. For a bunch of reasons, I'm also dubious about assuming that these cultures are representative of the lifestyles of pre-Industrial societies. Nonetheless, Siegel's participants apparently didn't have electricity or battery-powered devices. Anyway, Siegel found that sleep cycles changed over the seasons, but that their participants generally didn't have biphasic sleep, and didn't sleep much more than current people - 5.7-7.1 hours of sleep was the average. Siegel appears to believe that his data might mean that Ekirch's examples are specific to Northern Europe, with its much-colder temperature than you might get in Tanzania.

However, in a letter in the editor in the journal Sleep essentially replying to Siegel and colleagues, Ekirch points out some evidence of Europeans interacting with various non-European societies and referring to 'first sleep' and 'second sleep', including in Brazil in 1555, and Oman in the early 18th century. Ekirch also points out that some of his examples came from Spain, which is famously warmer in climate than England. Ekirch, finally, points out examples of native South American languages with different words for first and second sleep.

So there does appear to be historical and linguistic evidence of biphasic sleep occurring, though we can't be entirely sure of its prevalence, especially considering Siegel and colleagues' research in Tanzania, Namibia and Bolivia; the overall picture might be more complex than Ekirch is portraying it to be.