I think Natalia brings up a lot of important, true, points here, and I’m really glad she wrote this post. I also stand by what I said when I curated Alexey’s post, that while I was doubtful on any particular conclusion I thought he was pointing at real deficits in current models and I was excited to see those explored so they could be eventually filled in. My current guess (having not deeply investigated the evidence for either post) is that this post is closer to the literal truth, and Alexey’s post is pointing in a more useful direction (but we will get there in part by people pointing on the parts that are false, via posts like this).
The version of Alexey’s theses I most broadly support is “we treat sleep like one thing, when it is in fact multiple things with multiple purposes”. Natalia calls this epicycles, I agree with AllAmericanBreakfast that it’s more like scurvy before the vitamin C paradigm. There’s a real thing there, but without the paradigm it looks incredibly fake and exactly like epicycles. In general I think distinguishing ideas analogous to pre-vitamin-C scurvy from ideas like epicycles is one of the Hamming problems of science, and would cheer development in the area.
Some evidence I think is not explained by current mainstream sleep models:
Naps I have induced in myself via emotional or sensory overload, which have no relationship to how much I slept the night before or will sleep that night. I haven’t looked for studies on this but enough friends report similar things that I’m comfortable claiming the phenomenon definition exists.
Depression is sometimes cured through sleep deprivation.
Some people remain astonishingly productive for years on very little sleep.
Some evidence I think Alexey’s model fails to explain:
I couldn’t find information on the vitamin C density of polar bear livers in particular, but from these values, it seems far from clear that polar bear livers are more similar to lemons than limes in that respect. The vitamin C contents of limes and lime juice do not stand out in that list.
Moreover, it seems that it only takes about 10mg of vitamin C per day to prevent scurvy, and the Manual of Nutritional Therapeutics says that the same daily quantity is enough to improve scurvy symptoms, with 60-100mg/day being recommended for full recovery. So it seems that a cup (~240g) of fresh lime juice per day is enough to both prevent and enable recovery from scurvy.
(Scott and Scurvysays that “[t]ests on animals would later show that fresh lime juice has a quarter of the scurvy-fighting power of fresh lemon juice,” but I couldn’t find a source for that, I could find many contradicting it, and the USDA data suggests that it has ~3/4 of the vitamin C density of fresh lemon juice).
So it seems that the lime juice was just not preventing scurvy because it had spent long periods of time open to the air, and had been pumped through copper tubing. And so this paragraph from the post:
“Different kinds of citrus fruits are more like one another than they are like polar bear meat” sounds very reasonable, but in this case it was wrong. Sicilian lemons really ARE more like polar bear meat than they are like West Indian limes, at least for the purposes of treating scurvy.
seems very misleading (especially but not exclusively if you don’t interpret “polar bear meat” as referring to a specific polar bear organ, given that my sources say that fresh polar bear meat has 30x less vitamin C than fresh lime juice).
1. Copper tubing and environmental exposure does destroy vitamin C
So it seems that the lime juice was not preventing scurvy just because it had spent long periods of time open to the air, and had been pumped through copper tubing.
At first glance, I thought you were claiming that copper tubing and air exposure does not destroy vitamin C, but now I think I read you backwards. I think you are claiming that they do destroy vitamin C, and that copper and O2 destruction of vitamin C was the only reason that “lime juice” was failing to prevent scurvy (and that the differences between lemons vs. limes had nothing to do with it). Just to emphasize that vitamin C is susceptible to destruction by copper and environmental exposure, here’s the Australian Ministry of Health:
Vitamin C is very labile [easily broken down or displaced] and its content in foods varies. Vitamin C content can be affected by season, transport, shelf life, storage time, cooking practices and chlorination of water. Cutting, bruising, heating and exposure to copper, iron or mildly alkaline conditions can destroy ascorbate.
2. Liver and vitamin C
Polar bear liver contains massive amounts of vitamin A, enough to kill humans (1, 2). When the Idle Words essay on which the SMTM post is based says “Eat a bear liver every few weeks and scurvy will be the least of your problems,” the proper response is “because you’ll be dead of vitamin A poisoning.” However, the IW post vascillates between saying that the vitamin C in the polar bear comes from “meat,” “liver,” and “kidneys,” and I don’t trust that the author carefully investigated the nutritional value of each of these organs.
However, I did find a source that talks more about the Inuit and early European arctic explorer diet.
In the 1920s, the explorer Vilhjalmur Stefansson threw a wrench in “fresh fruits and vegetables” theory when he reported that the Inuit, who rarely ate plant foods, also did not suffer from scurvy. As it turns out, their Vitamin C came from animal organs: liver, adrenal glands, roe [egg masses in the ovaries], and tongue, to name a few. Traditional cooking methods also helped: Vitamin C is present in raw muscle meat, but harsh cooking destroys it. So by cooking their meat very lightly, the Inuit were preserving the Vitamin C in it. In a year-long experiment, Stefansson and a friend lived entirely on raw or lightly-cooked meat without showing any signs of scurvy.
I think we have adequate evidence to say that bear meat could have prevented scurvy in Arctic explorers if lightly cooked.
3. Sourcing the claims on limes vs. lemons
I found a source containing the claim of lemons being 4x as powerful an antiscorbutic as limes, which may be the origin of the IW/SMTM claim. In fact, it contains several of the essential points in the story conveyed by these blog posts. The source is The Prevention of Scurvy in the Navy by Surgeon-Commander J. L. PRISTON, M.B., M.R.C.P., D.P.H., R.N., from 1926. Unfortunately, neither SMTM nor Idle Words appear to cite this document.
In spite of this [better diet provided after a 1797 mutiny], scurvy continued to be a common disease for many years until, with the introduction of steam, voyages became shorter and fresh food was supplied as a matter of course, not as a rare privilege.
About the year 1860 the Admiralty gave up trying to get good supplies from the Mediterranean and began to use lime juice prepared from the sour lime in the West Indies. It is now known that the fresh lemon is four times as powerful an antiscorbutic as the fresh sour lime. Moreover, the method of preserving Navy lime juice destroys all the vitamin C, so that the Navy lime juice as now supplied has no antiscorbutic power at all.
The name “lime juice” had always been applied indiscriminately both to lemon juice and lime juice.
Arctic explorers, however, noticed from time to time as a puzzling fact that lime juice did not prevent scurvy.
We don’t really know what “fresh” means. Does it mean that both the limes and lemons were recently picked? Or only recently squeezed in the scientist’s lab, perhaps after several weeks in the ice-packed hull of a transport ship? And moreover, we don’t really know whether the “antiscorbutic power” of a citrus fruit scales linearly with vitamin C content, using whatever methods of processing and measurement they applied to generate this finding. Sadly, Priston also does not cite his source when he makes this claim.
Now let’s talk modern data. Is it possible that lemons contain, in some meaningful sense, about 4x the vitamin C of limes?
Key limes weigh about 1oz. Some random poster on a cooking forum says that a pound of key limes gives about 2⁄3 of a cup of juice, so you’d need 1.5 pounds = 24 oz of key limes to get a cup.
I’ve found varying estimates for how much juice is in a large lemon, but 3-5 tbsp seems to be about the range. Here’s a source for 5 tbsp. A 52mg/100g vitamin C lemon would have 1.7x the vitamin C content of an equivalent amount of lime juice, so you’d only need 0.6 cup, or 9.6 tbsp, to get the same amount. Rounding to 10, this would be the juice of two lemons, weighing in at about 8 oz.
By these figures, per unit weight of fresh fruit (as opposed to juice), fresh lemons might have 3x the vitamin C of fresh key limes. We can also imagine that if the scientist who conducted this old study was using lemons and limes purchased in their local market, those fruits might have been picked at different times. If the lemons were a bit fresher, perhaps more of the vitamin C was still stable. If Surgeon-General Priston was rounding up, or if random chance favored the lemon somewhat in the study, it’s possible to imagine a scientific result that at the time showed a pound of lemons being 4x as effective as a pound of key limes at preventing scurvy.
4. Conclusion
Note that these weird, subtle distinctions and epistemic challenges are precisely the point of the sleep/vitamin C comparison. Even if the early scientists studying the antiscorbutic power of lemons and limes wrongly found that lemons are 4x better than limes, this is precisely the point! Without both careful scientific investigation, using multiple lines of converging evidence (including the anecdotes of weirdos like arctic explorers), as well as a mechanistic understanding, it’s very easy for the real complexities of a system to be disguised by intuitive human categorization schemes that wildly miss the mark.
Another comparison is pica. We know that the body can generate cravings for “nutrients” that it does not need, or that are actively destructive (as in addiction or disordered eating), while also failing to give us cravings for nutrients that we do need. So we should keep our minds wide-open to the possibility that we are fundamentally misunderstanding the role of sleep in cognition and functional capacity.
At first glance, I thought you were claiming that copper tubing and air exposure does not destroy vitamin C, but now I think I read you backwards. I think you are claiming that they do destroy vitamin C, and that copper and O2 destruction of vitamin C was the only reason that “lime juice” was failing to prevent scurvy (and that the differences between lemons vs. limes had nothing to do with it).
Yeah, your current interpretation is correct.
This article finds that Musk limes have less than half the vitamin C of lemons, just to gesture in that direction.
It’s not that different from the raw limes in USDA’s data. In the titration method, that article found that musk limes had 42.3% as much vitamin C as lemons, and in the HPLC method, 53.5%. The average of those is 47.9%. My data on non-musk limes suggest they have 54.7% as much vitamin C as lemons.
The quote you gave for lemon vitamin C is the minimum value listed in the source. The maximum is 43.3mg/100g
I’m confused. The number I listed for lemon vitamin C was 53mg/100g.
I think we have adequate evidence to say that bear meat could have prevented scurvy in Arctic explorers if lightly cooked.
To be clear, I wasn’t arguing against that. I was just pointing out that it’s misleading to claim that polar bear meat is more like lemons than limes are. All 3 of those things seem able to prevent scurvy when they’re fresh and consumed in reasonable quantities; even 1⁄4 of the vitamin C concentration of fresh lemon juice is 23.22mg per cup. It’s possible that a citrus species has an abnormally low vitamin C concentration, and it’s possible that 10mg of vitamin C from lime juice is not the same as 10mg of vitamin C from fresh meat. But SMTM was asserting as a known fact that Sicilian lemons are more like polar bear meat than they are like West Indian limes in their ability to treat scurvy, which in expectation causes people who read the post to be worse at predicting the data we found than they would be if they had never read it, and that is what I was objecting to.
Apologies, I missed several of these points in your OP at first, and edited my post to correct it. Sorry you saw it before it was fixed!!!
I was just pointing out that it’s misleading to claim that polar bear meat is more like lemons than limes are.
I agree with your characterization of SMTM’s argument. It’s hard to parse, given the reference to his own made-up dialog that’s partly to make (supposedly) factual claims and partly to illustrate the challenges of analyzing complex scientific data. However, I read this as SMTM in their authorial voice defending the claim that key limes don’t contain vitamin C and polar bear meat does. We both agree this claim is false. Here’s the relevant quote:
Frolich:Maybe some citrus fruits contain the antiscorbutic [scurvy-curing] property and others don’t. Maybe the British Royal Navy used one kind of lime back when Lind did his research but gave a different kind of lime to Nares and the others on their Arctic expeditions. Or maybe they did something to the lime juice that removed the antiscorbutic property. Maybe they boiled it, or ran it through copper piping or something, and that ruined it.
Maybe there are different kinds of citrus. Maybe some animals need this mystery ingredient and others don’t. Maybe polar bear meat is, medically speaking, more like citrus fruit from Sicily than like citrus fruit from the West Indies. Really???
This looks a lot like special pleading, but in this case, the apparent double standard is correct. All of these weird exceptions he suggests were actually weird exceptions. And while our hypothetical version of Frolich wouldn’t have any way of knowing, these were the right distinctions to make.
I think that SMTM got this fact wrong. However, I think that the essential point of the SMTM article holds, and by extension the analogy I am drawing between vitamin C and sleep.
Thank you for your patience with my original slip-ups in noticing the full content of your preceding comment.
“Different kinds of citrus fruits are more like one another than they are like polar bear meat” sounds very reasonable, but in this case it was wrong. Sicilian lemons really ARE more like polar bear meat than they are like West Indian limes, at least for the purposes of treating scurvy.
Followup: SMTM points out that differences in fertilization and propagation (clonally or by seed) mean that we don’t know whether or not historical key limes had enough vitamin C to cure scurvy. I notice in myself a certain tendency to say oh puh-leeze in response to this, but at the same time, I think the broader point of the original SMTM piece is that it’s precisely this reaction that we ought to be suspicious of. After all, we don’t necessarily know that the arctic explorers who found that lime juice didn’t cure their scurvy were drinking it out of copper tubes, or that the vitamin C in the lime juice they were drinking was oxidized.
SMTM points out that differences in fertilization and propagation (clonally or by seed) mean that we don’t know whether or not historical key limes had enough vitamin C to cure scurvy.
Hm, that would still make the post misleading — “Sicilian lemons really ARE more like polar bear meat than they are like West Indian limes, at least for the purposes of treating scurvy” is a very different claim from “we don’t know whether or not historical key limes had enough vitamin C to cure scurvy.”
The version of Alexey’s theses I most broadly support is “we treat sleep like one thing, when it is in fact multiple things with multiple purposes”. Natalia calls this epicycles
That sleep is multiple things is not something I am arguing against. (In fact, I don’t recall that being a point in Guzey’s post). What I uncharitably called “epicycles” was the additional complexity Guzey’s model has to have to explain why so many people feel dumber after sleep restriction, and why experimental studies say that sleep loss causes cognitive impairments, when ““not sleeping ‘enough’ makes you stupid” is a 100% psyop.”
It’s pretty clear to me that sleep has multiple effects, and that it might be the case that there’s something with all of sleep’s good effects and none of the bad ones that just hasn’t been discovered yet. Maybe digital people would only have to spend the subjective equivalent of a few seconds per day shut down in order to renormalize their weights or whatever, or might not need to be shut down ever at all to maintain their performance.
But I don’t think that is incompatible with the object-level claims in my post, any more than saying “humans don’t live to be 200 years old” is incompatible with knowing that the passage of time has multiple effects or with admitting the possibility of revolutionary life-extension methods being discovered that do allow humans to be 200 years old.
Sounds like we agree a lot- I think he got a lot of stuff wrong and was glad to see you correct it. I think he wildly overextrapolated from his own data. My opinion on mainstream sleep researchers is closer to his than yours, but agree he overextrapolated from the sailing sleep coach.
I spend a lot of time going “that person’s big idea is built on a foundation of factually inaccurate sand”, so it’s novel for me to be the one going “yes, mostly, but I still want to follow up”. I feel this way due to a combination of thinking Alexey brings up a lot of legit problems with existing sleep research, and that he put serious time into generating new data (although I was extremely disappointed in his data collection, and would have found even writing down his subjective score each day an improvement over nothing).
My opinion on mainstream sleep researchers is closer to his than yours
I feel this way due to a combination of thinking Alexey brings up a lot of legit problems with existing sleep research
Could you elaborate on what you mean by “mainstream sleep researchers” and/or “existing sleep research,” and what, in your view, the legit problems with the latter are? (Modulo Walker). These terms can mean many things and refer to very different sets of models. Researchers, the organizations that set guidelines, and clinicians in a certain field can often have very mistaken views about what the research in their field actually shows if they’re not personally involved in it. For example, some clinical psychiatrists promote the serotonin theory of depression even though the research body does not support it. Lumping together the clinicians with the research body as a monolithic “psychiatric field” obscures this. (I’m not sure that that’s what’s going on here but discussing what you mean specifically could be helpful.)
Some of Guzey’s points on Theses on Sleep and Matthew Walker’s “Why We Sleep” Is Riddled with Scientific and Factual Errors are genuinely both counterintuitive to the general public (and some/most sleep researchers, guidelines-setters, and clinicians) and true. But Guzey’s evidence for those points didn’t come out of the ether (and most of it did not come from his own experiment). It largely came out of the academic research body on sleep, the same place in which I found evidence that sleep duration e.g. affects cognition but barely has a clear association with mortality. So it seems reasonable to me to dissociate the research body from the opinions of most people in the field.
(Also, I think it’s very unclear what view you’re ascribing to me in the first quote, but I suspect that I do not endorse it. I don’t expect sleep scientists to be that good at predicting reality in subfields or sub-subfields they’re not personally involved in.)
In general I think distinguishing ideas analogous to pre-vitamin-C scurvy from ideas like epicycles is one of the Hamming problems of science, and would cheer development in the area.
I’ll take a stab at this.
It seems like the key distinction is separating ideas that are heavily but still inadequately tested (epicycles) from those that are underexplored (pre-vitamin-C scurvy) and those that are thoroughly tested (the pathogen theory of disease).
Researchers are bound by convention, resources, and the limits of their own imagination and enthusiasm. As such, they often wind up using experimental methods that both fail to distinguish empirical confounders and can never yield a better underlying mechanistic account of the phenomena at hand. Instead, they just put out another study as meta-analysis fodder.
Take this sleep research. We can find dozens of studies showing that X hours of sleep deprivation results in Y amount of cognitive deficit. But have the scientists designed studies to test for weird confounds?
Maybe being sleepy doesn’t make you cognitively slow, but instead makes you socially anxious. Being around a bunch of strange scientists subjecting you to unusual tests in an unfamiliar environment could trigger anxiety. This in turn could explain some of the cognitive deficits in the sleep research. Sleepiness → social anxiety → cognitive deficits around strangers. That hypothesis took me all of 1 minute to come up with, and it might be true, who knows? If resources were unlimited for testing it, we could just go ahead and do it.
If it takes me 1 minute to come up with a probably-untested plausible alternative hypothesis to a field that has produced dozens and dozens of studies, it’s a sign that the field is designing its experiments inefficiently, and we should discount the apparent strength of its findings. This doesn’t mean that its current hypothesis is wrong, but that it has disguished alternative hypotheses by neglecting to test them and thus created a false impression of consensus.
If it’s not possible to test all the hypotheses we’d ideally like to explore, then too bad. Scientists don’t get to say they’ve discovered “the truth” just because they’ve hit the current limit of their experimental capacities. They just have to admit that they’re not sure. In this case, we would start by pointing out that we have extremely limited mechanistic information on exactly how sleep and extended wakefulness physically alter the brain, and how any such alterations turn into cognitive impacts. By contrast, we have abundant (if still far from completely sufficient) physical information about how other complex physiological systems work, such as the immune system.
So in general, we ought to rank our certainty not by how much evidence we’ve accumulated, but by how many alternative hypotheses we’ve ruled out (this idea is certainly not original to me). Unfortunately, either there’s no good way to quantify this, or there is, and it’s just not conventional for scientists to do so when designing their experiments and writing their literature reviews.
Based on my limited experience in the lab environment, I can already see firsthand the degree to which resource constraints, convention, careerism, and an honest assessment of one’s own strengths and limitations as a scientist dominate decisions about what to study and how to study it. “Important if true” gives way to other concerns, like “who will pay to find out?.”
I think science tends to get the biggest big things right, and that’s a profound accomplishment. But if it’s a question with significant nuance, in a niche subfield, and doesn’t have any really crisp methods and unequivocal metrics, then you’re at the frontier, and should retain durable skepticism about the field.
I don’t think that’s much of a problem. People get “hangry” too and that doesn’t invalidate the usefulness of fasting nor does it imply that fasting has to be miserable like that.
I think any prolonged discomfort that is unaddressed is going to make people cranky and bias them towards negative explanations of things. Address the discomfort, and the experience can change dramatically.
Perhaps the most useful thing I got out of experimenting with polyphasic sleep was a recognition of how well I could function when sleep deprived so long as I wasn’t constantly battling an unaddressed urge to sleep. Sleep deprivation still made me dumber, but the crankiness and most of the dysfunction was actually a result of trying to ignore (and yet not completely tuning out) my body screaming at me to sleep. Having a reference experience for being sleep deprived without craving sleep made it a lot easier to function when an hour or two short, and also more willing to go take a nap when I need one.
Some evidence I think is not explained by current mainstream sleep models:
Naps I have induced in myself via emotional or sensory overload, which have no relationship to how much I slept the night before or will sleep that night. I haven’t looked for studies on this but enough friends report similar things that I’m comfortable claiming the phenomenon definition exists.
Depression is sometimes cured through sleep deprivation.
Some people remain astonishingly productive for years on very little sleep.
I don’t think the last point is not explained by mainstream sleep models. Around 440,000 people in the United States are 3 standard deviations smarter than the US average. That, together with the fact that some jobs (even some that are common among those 440,000 people) might have higher marginal returns to hours worked & past experience than to cognitive performance, makes me unsurprised that some people remain astonishingly productive for years on very little sleep.
I’m not sure what would count as “mainstream sleep models” w.r.t sleep deprivation’s antidepressive effects, but a lot of studies seem to confirm that acute SD increases BDNF, as Guzey pointed out.
I think mainstream sleep models don’t fail to predict that some people remain astonishingly productive for years on very little sleep. The meta-analyses I included showed moderate effect sizes for the impacts of sleep restriction on cognition, which aren’t enough to decrease a large baseline advantage in cognitive performance that much.
okay. It sounds like the words “sleep deprivation” are doing a lot of work in the original post. You don’t mean “sleeping much less than the average”, but literally “depriving yourself of sleep you need”?
Regardless of what mainstream science says, what gets filtered out to the public is very much “get N hours or you’re going to die”, if sleep scientists have the concept of standard deviation in their models I’ll be impressed.
Let’s say that long-term sleep deprivation costs 15 IQ points (1 standard deviation) in terms of cognitive function.
Furthermore, let’s imagine Bob is a researcher with a baseline IQ of 145, while his coworkers have IQs of 115. Bob is chronically sleep deprived, so his functioning IQ is 130.
Bob’s functioning IQ is still above that of his coworkers, and he’s also able to work more hours, and therefore will be astonishinly productive relative to his coworkers.
However, it may still be that his productivity would be higher still if he slept better. What he lost in time, he’d more than make up for in productivity per hour of work.
Therefore, the existence of rare, astonishingly productive, sleep-deprived workers is compatible with sleep deprivation being harmful to cognitive function and net negative to productivity even for these workers.
Let’s assume that at a research university, the mean IQ of faculty is 115, still with a standard deviation of 15 IQ points. In this case, Bob would merely be in the top 5% for IQ relative to his colleagues. At a major research university, with 5,000 faculty, there would be 250 faculty like Bob. We can imagine that their level of sleep deprivation would be normally distributed as well, and so we could find quite a few examples of people like Bob just by looking around.
Natália, is that a fair representation of the point you’re making here?
I think that makes sense (and wasn’t what Natalia was saying, but is compatible with it. Bob could be smater than average, have a lower optimal sleep length, and have fewer consequences per hour of foregone sleep ). I want to highlight that there are many tasks that don’t benefit from the additional 15 IQ points, and if your total output depends more on those than on tasks that benefit from every last IQ point, then trading IQ points for more hours in the day can be the optimal move. But if Bob got smarter co-workers or became less resilient to sleep loss the trade-off might be different, and it’s beneficial for him to check the trade-offs periodicially.
I don’t think that’s what Alexey or Natalia were arguing, and it ignores health effects, but what I’m excited about in this discussion is rendering these kinds of trade-offs able to reasoned about
I agree about your point on making trade-offs possible to reason about. I’ve found that in discussing this away from LW, people will talk as if extra hours of sleep don’t come at the expense of waking hours. To them, at first glance, the decision is between “activity X, Y, and Z with 6 or 8 hours of sleep” rather than between “activity X and Y with 8 hours of sleep, or activity X, Y, and Z with 6 hours of sleep.”
When I point this out, the follow-up is “but what if the lost sleep makes activity Z a negative experience?” When we eventually anchor on that being the question at hand, rather than the assumption by default, the conversation can become productive and interesting.
I also agree with your bottlenecks-based framing. I think some tasks are bottlenecked by cognitive function (chess competitions), others usually aren’t but can be in rare/emergency circumstances that therefore require full-time high cognitive function (airline pilots), and many others aren’t bottlenecked by cognition (sailing races).
One counterpoint to that framing is that activities usually don’t take a fixed amount of time. There are many things that I’ve done where I’ll take a quarter of the time or less if I’m in a productive state rather than a non-productive state. Sometimes it might really be a trade off between “activity X, Y, and Z with 6 or 8 hours of sleep”, or even “activity X, Y, and Z with 8 hours of sleep, or activity X and Y with 6 hours of sleep”. I’ve also found that I’m less distracted with more sleep, so it could be “be productive for the most of the day with 8 hours of sleep, or browse social media for most of the day with 6 hours of sleep”. This all has been most obvious when I’ve been working on intellectual activities, like writing or developing software, but I also tend to have slower reflexes when I’ve been sleeping less, so physical activities are also relevant.
My personal sleep logs show that productive states are strongly correlated with the amount of hours of sleep I’ve gotten over the last month and the previous night. I don’t have proof that the math really works out, but I strongly suspect it does. Over longer term periods where I’ve had more or fewer hours of sleep, I’ve gotten more positive outside feedback (grades, feedback in monthly reviews) in the the periods with more sleep. But of course I have no proof that this applies to everyone, it’s correlation instead of causation, and it’s certainly no double blind study.
And of course the hardest problem is beating the akrasia to go to bed when I want to, but that’s a separate issue.
Agree! I think we are coming to a model where N hours of SD results in an average of X% loss in productivity per hour, and a change from Y to Z daily productive hours. We want to know the sign of XZ—Y.
Here, you may be arguing both that the performance loss is so great as to more than cancel out the increased number of working hours. You may also be arguing that we actually lose productive hours with sleep deprivation, despite being awake longer, because the loss of executive function makes us unable to focus on our work.
Alternatively, it may be that the additional productive hours from SD can be put into low-demand but time-consuming life maintenance or pleasurable occupations that otherwise might get neglected. For example, we can imagine a busy student choosing between 8 hours of sleep, or 6 hours of sleep plus 2 hours of socializing. Or between 8 hours of sleep, or 7 hours of sleep plus a trip to the grocery store. Or between 8 hours of sleep and 7 hours of sleep plus 1 hour of exercise.
Sacrificing sleep for efficiency-compounding tasks and meaningful occupations may cancel out or even reverse the cognitive impacts of SD itself. It may be that the worst effects of SD can be mitigated with a combination of practice, organizational supports, and stimulating activities and caffeine. It may also be that for some people, their baseline need for sleep is below the level recommended by sleep researchers, so that for them, reduced sleep comes at no cognitive cost.
So the question is twofold. Does non-strategic mild-to-moderate SD lead to net positive or net negative results on some outcome, like productivity, pleasure, health, or meaningful experiences? Furthermore, to what extent can an SD-management strategy mitigate the negative impacts of mild SD?
The stakes are rather significant. Consider that 2 hours of daily SD from age 20-60 adds up to 3.3 years of additional wakefulness, distributed across the most productive and healthy part of the lifespan. If it ruins the experience of life and damages productivity, that is an enormous detriment. On the other hand, if long-term strategically managed SD can be made neutral to positive, then that is the only intervention we currently have to more or less add healthy years to the natural lifespan.
Given the obvious complexities of managing sleep, even for adults people who are attempting to get a full 7-8 hours per night, it seems worth puzzling through it up front in order to establish an optimal sleep routine, given one’s individual preferences, bodily needs, and demands of life.
It sounds like the words “sleep deprivation” are doing a lot of work in the original post. You don’t mean “sleeping much less than the average”, but literally “depriving yourself of sleep you need”?
I tried to be explicit about the average number of hours slept in the studies included in these meta-analyses of experimental studies. Usually, it’s about 4 to 5 hours, with the average seeming to be closer to 4. I also tried to be explicit about how many hours of sleep per night my claims refer to in my conclusion (except for the first claim, which I specified was about “shorter-than-average” sleep, and the last one, which is a harder-to-operationalize claim, but still didn’t explicitly refer to “sleep deprivation”).
if sleep scientists have the concept of standard deviation in their models I’ll be impressed.
Standard deviations are used very often; I would be surprised if they weren’t used in 5% or more of the sleep studies I’ve read.
Sorry yes, of course they’re including SD calculations in their papers. What I meant was that medical papers (or popular coverage thereof) almost never consider the implications of the fact that 0.3% of people will be >=3 standard devs out, and with a large enough population that’s a lot of people.
I think Natalia brings up a lot of important, true, points here, and I’m really glad she wrote this post. I also stand by what I said when I curated Alexey’s post, that while I was doubtful on any particular conclusion I thought he was pointing at real deficits in current models and I was excited to see those explored so they could be eventually filled in. My current guess (having not deeply investigated the evidence for either post) is that this post is closer to the literal truth, and Alexey’s post is pointing in a more useful direction (but we will get there in part by people pointing on the parts that are false, via posts like this).
The version of Alexey’s theses I most broadly support is “we treat sleep like one thing, when it is in fact multiple things with multiple purposes”. Natalia calls this epicycles, I agree with AllAmericanBreakfast that it’s more like scurvy before the vitamin C paradigm. There’s a real thing there, but without the paradigm it looks incredibly fake and exactly like epicycles. In general I think distinguishing ideas analogous to pre-vitamin-C scurvy from ideas like epicycles is one of the Hamming problems of science, and would cheer development in the area.
Some evidence I think is not explained by current mainstream sleep models:
Naps I have induced in myself via emotional or sensory overload, which have no relationship to how much I slept the night before or will sleep that night. I haven’t looked for studies on this but enough friends report similar things that I’m comfortable claiming the phenomenon definition exists.
Depression is sometimes cured through sleep deprivation.
Some people remain astonishingly productive for years on very little sleep.
Some evidence I think Alexey’s model fails to explain:
“do all my friends hate me or do I just need a nap”
Or more formally, the evidence on depression Natalia links
Alexey acknowledges that sleep deprivation makes people much worse drivers but doesn’t seem to think that has implications for anything else?
In a 1:1 conversation with me he argued that medication never increases the actual need for sleep, merely perceived need.
[ETA: to be clear, I’m not criticizing the thesis of SMTM’s post here, just pointing out a factual error]
The linked SMTM post is misleading.
Here is the vitamin C content per 100g of some relevant foods, which I found after a few minutes of searching on Google:
Lime juice: 30mg [1]
Lemon juice: 38.7mg [2]
Raw lemons: 53mg [3]
Raw limes: 29mg [4]
Key limes in particular, which Wikipedia says are the same thing as West Indian limes: 31.3mg [5]
Raw caribou liver: 23.874.9 mg [6]
Raw ringed seal liver: 23.873.8 mg [6]
Raw cattle liver: 71.2mg [7]
Raw buffalo liver: 72.4mg [7]
Raw sheep liver: 77.6mg [7]
Raw goat liver: 76.7mg [7]
Fresh/raw polar bear meat: 1mg [8] [9]
I couldn’t find information on the vitamin C density of polar bear livers in particular, but from these values, it seems far from clear that polar bear livers are more similar to lemons than limes in that respect. The vitamin C contents of limes and lime juice do not stand out in that list.
Moreover, it seems that it only takes about 10mg of vitamin C per day to prevent scurvy, and the Manual of Nutritional Therapeutics says that the same daily quantity is enough to improve scurvy symptoms, with 60-100mg/day being recommended for full recovery. So it seems that a cup (~240g) of fresh lime juice per day is enough to both prevent and enable recovery from scurvy.
(Scott and Scurvy says that “[t]ests on animals would later show that fresh lime juice has a quarter of the scurvy-fighting power of fresh lemon juice,” but I couldn’t find a source for that, I could find many contradicting it, and the USDA data suggests that it has ~3/4 of the vitamin C density of fresh lemon juice).
So it seems that the lime juice was just not preventing scurvy because it had spent long periods of time open to the air, and had been pumped through copper tubing. And so this paragraph from the post:
seems very misleading (especially but not exclusively if you don’t interpret “polar bear meat” as referring to a specific polar bear organ, given that my sources say that fresh polar bear meat has 30x less vitamin C than fresh lime juice).
https://fdc.nal.usda.gov/fdc-app.html#/food-details/168156/nutrients
https://fdc.nal.usda.gov/fdc-app.html#/food-details/167747/nutrients
https://fdc.nal.usda.gov/fdc-app.html#/food-details/1102594/nutrients
https://fdc.nal.usda.gov/fdc-app.html#/food-details/168155/nutrients
https://fdc.nal.usda.gov/fdc-app.html#/food-details/509702/nutrients
https://doi.org/10.1006/jfca.2002.1053
https://doi.org/10.1111/j.1365-2621.1965.tb00287.x
https://www.jstor.org/stable/40508955
https://escholarship.mcgill.ca/downloads/t148fj883?locale=en
1. Copper tubing and environmental exposure does destroy vitamin C
At first glance, I thought you were claiming that copper tubing and air exposure does not destroy vitamin C, but now I think I read you backwards. I think you are claiming that they do destroy vitamin C, and that copper and O2 destruction of vitamin C was the only reason that “lime juice” was failing to prevent scurvy (and that the differences between lemons vs. limes had nothing to do with it). Just to emphasize that vitamin C is susceptible to destruction by copper and environmental exposure, here’s the Australian Ministry of Health:
2. Liver and vitamin C
Polar bear liver contains massive amounts of vitamin A, enough to kill humans (1, 2). When the Idle Words essay on which the SMTM post is based says “Eat a bear liver every few weeks and scurvy will be the least of your problems,” the proper response is “because you’ll be dead of vitamin A poisoning.” However, the IW post vascillates between saying that the vitamin C in the polar bear comes from “meat,” “liver,” and “kidneys,” and I don’t trust that the author carefully investigated the nutritional value of each of these organs.
However, I did find a source that talks more about the Inuit and early European arctic explorer diet.
Raw bear meat has a little more vitamin C (2mg/100g). So this would require eating 500g bear meat/day (about 650 calories, or 1.1lb per day) to get an adequate quantity to prevent scurvy indefinitely, or less merely to prevent it for the duration of the voyage. Scurvy takes 3 months to set in if vitamin C intake is extremely low.
I think we have adequate evidence to say that bear meat could have prevented scurvy in Arctic explorers if lightly cooked.
3. Sourcing the claims on limes vs. lemons
I found a source containing the claim of lemons being 4x as powerful an antiscorbutic as limes, which may be the origin of the IW/SMTM claim. In fact, it contains several of the essential points in the story conveyed by these blog posts. The source is The Prevention of Scurvy in the Navy by Surgeon-Commander J. L. PRISTON, M.B., M.R.C.P., D.P.H., R.N., from 1926. Unfortunately, neither SMTM nor Idle Words appear to cite this document.
We don’t really know what “fresh” means. Does it mean that both the limes and lemons were recently picked? Or only recently squeezed in the scientist’s lab, perhaps after several weeks in the ice-packed hull of a transport ship? And moreover, we don’t really know whether the “antiscorbutic power” of a citrus fruit scales linearly with vitamin C content, using whatever methods of processing and measurement they applied to generate this finding. Sadly, Priston also does not cite his source when he makes this claim.
Now let’s talk modern data. Is it possible that lemons contain, in some meaningful sense, about 4x the vitamin C of limes?
Key limes weigh about 1oz. Some random poster on a cooking forum says that a pound of key limes gives about 2⁄3 of a cup of juice, so you’d need 1.5 pounds = 24 oz of key limes to get a cup.
I’ve found varying estimates for how much juice is in a large lemon, but 3-5 tbsp seems to be about the range. Here’s a source for 5 tbsp. A 52mg/100g vitamin C lemon would have 1.7x the vitamin C content of an equivalent amount of lime juice, so you’d only need 0.6 cup, or 9.6 tbsp, to get the same amount. Rounding to 10, this would be the juice of two lemons, weighing in at about 8 oz.
By these figures, per unit weight of fresh fruit (as opposed to juice), fresh lemons might have 3x the vitamin C of fresh key limes. We can also imagine that if the scientist who conducted this old study was using lemons and limes purchased in their local market, those fruits might have been picked at different times. If the lemons were a bit fresher, perhaps more of the vitamin C was still stable. If Surgeon-General Priston was rounding up, or if random chance favored the lemon somewhat in the study, it’s possible to imagine a scientific result that at the time showed a pound of lemons being 4x as effective as a pound of key limes at preventing scurvy.
4. Conclusion
Note that these weird, subtle distinctions and epistemic challenges are precisely the point of the sleep/vitamin C comparison. Even if the early scientists studying the antiscorbutic power of lemons and limes wrongly found that lemons are 4x better than limes, this is precisely the point! Without both careful scientific investigation, using multiple lines of converging evidence (including the anecdotes of weirdos like arctic explorers), as well as a mechanistic understanding, it’s very easy for the real complexities of a system to be disguised by intuitive human categorization schemes that wildly miss the mark.
Another comparison is pica. We know that the body can generate cravings for “nutrients” that it does not need, or that are actively destructive (as in addiction or disordered eating), while also failing to give us cravings for nutrients that we do need. So we should keep our minds wide-open to the possibility that we are fundamentally misunderstanding the role of sleep in cognition and functional capacity.
Yeah, your current interpretation is correct.
It’s not that different from the raw limes in USDA’s data. In the titration method, that article found that musk limes had 42.3% as much vitamin C as lemons, and in the HPLC method, 53.5%. The average of those is 47.9%. My data on non-musk limes suggest they have 54.7% as much vitamin C as lemons.
I’m confused. The number I listed for lemon vitamin C was 53mg/100g.
To be clear, I wasn’t arguing against that. I was just pointing out that it’s misleading to claim that polar bear meat is more like lemons than limes are. All 3 of those things seem able to prevent scurvy when they’re fresh and consumed in reasonable quantities; even 1⁄4 of the vitamin C concentration of fresh lemon juice is 23.22mg per cup. It’s possible that a citrus species has an abnormally low vitamin C concentration, and it’s possible that 10mg of vitamin C from lime juice is not the same as 10mg of vitamin C from fresh meat. But SMTM was asserting as a known fact that Sicilian lemons are more like polar bear meat than they are like West Indian limes in their ability to treat scurvy, which in expectation causes people who read the post to be worse at predicting the data we found than they would be if they had never read it, and that is what I was objecting to.
Apologies, I missed several of these points in your OP at first, and edited my post to correct it. Sorry you saw it before it was fixed!!!
I agree with your characterization of SMTM’s argument. It’s hard to parse, given the reference to his own made-up dialog that’s partly to make (supposedly) factual claims and partly to illustrate the challenges of analyzing complex scientific data. However, I read this as SMTM in their authorial voice defending the claim that key limes don’t contain vitamin C and polar bear meat does. We both agree this claim is false. Here’s the relevant quote:
I think that SMTM got this fact wrong. However, I think that the essential point of the SMTM article holds, and by extension the analogy I am drawing between vitamin C and sleep.
Thank you for your patience with my original slip-ups in noticing the full content of your preceding comment.
There was also another relevant passage:
Followup: SMTM points out that differences in fertilization and propagation (clonally or by seed) mean that we don’t know whether or not historical key limes had enough vitamin C to cure scurvy. I notice in myself a certain tendency to say oh puh-leeze in response to this, but at the same time, I think the broader point of the original SMTM piece is that it’s precisely this reaction that we ought to be suspicious of. After all, we don’t necessarily know that the arctic explorers who found that lime juice didn’t cure their scurvy were drinking it out of copper tubes, or that the vitamin C in the lime juice they were drinking was oxidized.
Reality is surprisingly complex.
Hm, that would still make the post misleading — “Sicilian lemons really ARE more like polar bear meat than they are like West Indian limes, at least for the purposes of treating scurvy” is a very different claim from “we don’t know whether or not historical key limes had enough vitamin C to cure scurvy.”
Thanks for contacting them, though!
Agreed!
I will see if I can get in touch with SMTM and IW, I’d be curious what they have to say about this.
That sleep is multiple things is not something I am arguing against. (In fact, I don’t recall that being a point in Guzey’s post). What I uncharitably called “epicycles” was the additional complexity Guzey’s model has to have to explain why so many people feel dumber after sleep restriction, and why experimental studies say that sleep loss causes cognitive impairments, when ““not sleeping ‘enough’ makes you stupid” is a 100% psyop.”
It’s pretty clear to me that sleep has multiple effects, and that it might be the case that there’s something with all of sleep’s good effects and none of the bad ones that just hasn’t been discovered yet. Maybe digital people would only have to spend the subjective equivalent of a few seconds per day shut down in order to renormalize their weights or whatever, or might not need to be shut down ever at all to maintain their performance.
But I don’t think that is incompatible with the object-level claims in my post, any more than saying “humans don’t live to be 200 years old” is incompatible with knowing that the passage of time has multiple effects or with admitting the possibility of revolutionary life-extension methods being discovered that do allow humans to be 200 years old.
Sounds like we agree a lot- I think he got a lot of stuff wrong and was glad to see you correct it. I think he wildly overextrapolated from his own data. My opinion on mainstream sleep researchers is closer to his than yours, but agree he overextrapolated from the sailing sleep coach.
I spend a lot of time going “that person’s big idea is built on a foundation of factually inaccurate sand”, so it’s novel for me to be the one going “yes, mostly, but I still want to follow up”. I feel this way due to a combination of thinking Alexey brings up a lot of legit problems with existing sleep research, and that he put serious time into generating new data (although I was extremely disappointed in his data collection, and would have found even writing down his subjective score each day an improvement over nothing).
Could you elaborate on what you mean by “mainstream sleep researchers” and/or “existing sleep research,” and what, in your view, the legit problems with the latter are? (Modulo Walker). These terms can mean many things and refer to very different sets of models. Researchers, the organizations that set guidelines, and clinicians in a certain field can often have very mistaken views about what the research in their field actually shows if they’re not personally involved in it. For example, some clinical psychiatrists promote the serotonin theory of depression even though the research body does not support it. Lumping together the clinicians with the research body as a monolithic “psychiatric field” obscures this. (I’m not sure that that’s what’s going on here but discussing what you mean specifically could be helpful.)
Some of Guzey’s points on Theses on Sleep and Matthew Walker’s “Why We Sleep” Is Riddled with Scientific and Factual Errors are genuinely both counterintuitive to the general public (and some/most sleep researchers, guidelines-setters, and clinicians) and true. But Guzey’s evidence for those points didn’t come out of the ether (and most of it did not come from his own experiment). It largely came out of the academic research body on sleep, the same place in which I found evidence that sleep duration e.g. affects cognition but barely has a clear association with mortality. So it seems reasonable to me to dissociate the research body from the opinions of most people in the field.
(Also, I think it’s very unclear what view you’re ascribing to me in the first quote, but I suspect that I do not endorse it. I don’t expect sleep scientists to be that good at predicting reality in subfields or sub-subfields they’re not personally involved in.)
[An earlier version of this comment said “do endorse it” when I meant the opposite, sorry for the confusion.]
I’ll take a stab at this.
It seems like the key distinction is separating ideas that are heavily but still inadequately tested (epicycles) from those that are underexplored (pre-vitamin-C scurvy) and those that are thoroughly tested (the pathogen theory of disease).
Researchers are bound by convention, resources, and the limits of their own imagination and enthusiasm. As such, they often wind up using experimental methods that both fail to distinguish empirical confounders and can never yield a better underlying mechanistic account of the phenomena at hand. Instead, they just put out another study as meta-analysis fodder.
Take this sleep research. We can find dozens of studies showing that X hours of sleep deprivation results in Y amount of cognitive deficit. But have the scientists designed studies to test for weird confounds?
Maybe being sleepy doesn’t make you cognitively slow, but instead makes you socially anxious. Being around a bunch of strange scientists subjecting you to unusual tests in an unfamiliar environment could trigger anxiety. This in turn could explain some of the cognitive deficits in the sleep research. Sleepiness → social anxiety → cognitive deficits around strangers. That hypothesis took me all of 1 minute to come up with, and it might be true, who knows? If resources were unlimited for testing it, we could just go ahead and do it.
If it takes me 1 minute to come up with a probably-untested plausible alternative hypothesis to a field that has produced dozens and dozens of studies, it’s a sign that the field is designing its experiments inefficiently, and we should discount the apparent strength of its findings. This doesn’t mean that its current hypothesis is wrong, but that it has disguished alternative hypotheses by neglecting to test them and thus created a false impression of consensus.
If it’s not possible to test all the hypotheses we’d ideally like to explore, then too bad. Scientists don’t get to say they’ve discovered “the truth” just because they’ve hit the current limit of their experimental capacities. They just have to admit that they’re not sure. In this case, we would start by pointing out that we have extremely limited mechanistic information on exactly how sleep and extended wakefulness physically alter the brain, and how any such alterations turn into cognitive impacts. By contrast, we have abundant (if still far from completely sufficient) physical information about how other complex physiological systems work, such as the immune system.
So in general, we ought to rank our certainty not by how much evidence we’ve accumulated, but by how many alternative hypotheses we’ve ruled out (this idea is certainly not original to me). Unfortunately, either there’s no good way to quantify this, or there is, and it’s just not conventional for scientists to do so when designing their experiments and writing their literature reviews.
Based on my limited experience in the lab environment, I can already see firsthand the degree to which resource constraints, convention, careerism, and an honest assessment of one’s own strengths and limitations as a scientist dominate decisions about what to study and how to study it. “Important if true” gives way to other concerns, like “who will pay to find out?.”
I think science tends to get the biggest big things right, and that’s a profound accomplishment. But if it’s a question with significant nuance, in a niche subfield, and doesn’t have any really crisp methods and unequivocal metrics, then you’re at the frontier, and should retain durable skepticism about the field.
I don’t think that’s much of a problem. People get “hangry” too and that doesn’t invalidate the usefulness of fasting nor does it imply that fasting has to be miserable like that.
I think any prolonged discomfort that is unaddressed is going to make people cranky and bias them towards negative explanations of things. Address the discomfort, and the experience can change dramatically.
Perhaps the most useful thing I got out of experimenting with polyphasic sleep was a recognition of how well I could function when sleep deprived so long as I wasn’t constantly battling an unaddressed urge to sleep. Sleep deprivation still made me dumber, but the crankiness and most of the dysfunction was actually a result of trying to ignore (and yet not completely tuning out) my body screaming at me to sleep. Having a reference experience for being sleep deprived without craving sleep made it a lot easier to function when an hour or two short, and also more willing to go take a nap when I need one.
I don’t think the last point is not explained by mainstream sleep models. Around 440,000 people in the United States are 3 standard deviations smarter than the US average. That, together with the fact that some jobs (even some that are common among those 440,000 people) might have higher marginal returns to hours worked & past experience than to cognitive performance, makes me unsurprised that some people remain astonishingly productive for years on very little sleep.
I’m not sure what would count as “mainstream sleep models” w.r.t sleep deprivation’s antidepressive effects, but a lot of studies seem to confirm that acute SD increases BDNF, as Guzey pointed out.
I’m surprised to hear you say this and am now less sure what you’re arguing against.
I think mainstream sleep models don’t fail to predict that some people remain astonishingly productive for years on very little sleep. The meta-analyses I included showed moderate effect sizes for the impacts of sleep restriction on cognition, which aren’t enough to decrease a large baseline advantage in cognitive performance that much.
okay. It sounds like the words “sleep deprivation” are doing a lot of work in the original post. You don’t mean “sleeping much less than the average”, but literally “depriving yourself of sleep you need”?
Regardless of what mainstream science says, what gets filtered out to the public is very much “get N hours or you’re going to die”, if sleep scientists have the concept of standard deviation in their models I’ll be impressed.
My interpretation is something like this:
Let’s say that long-term sleep deprivation costs 15 IQ points (1 standard deviation) in terms of cognitive function.
Furthermore, let’s imagine Bob is a researcher with a baseline IQ of 145, while his coworkers have IQs of 115. Bob is chronically sleep deprived, so his functioning IQ is 130.
Bob’s functioning IQ is still above that of his coworkers, and he’s also able to work more hours, and therefore will be astonishinly productive relative to his coworkers.
However, it may still be that his productivity would be higher still if he slept better. What he lost in time, he’d more than make up for in productivity per hour of work.
Therefore, the existence of rare, astonishingly productive, sleep-deprived workers is compatible with sleep deprivation being harmful to cognitive function and net negative to productivity even for these workers.
Let’s assume that at a research university, the mean IQ of faculty is 115, still with a standard deviation of 15 IQ points. In this case, Bob would merely be in the top 5% for IQ relative to his colleagues. At a major research university, with 5,000 faculty, there would be 250 faculty like Bob. We can imagine that their level of sleep deprivation would be normally distributed as well, and so we could find quite a few examples of people like Bob just by looking around.
Natália, is that a fair representation of the point you’re making here?
I think that makes sense (and wasn’t what Natalia was saying, but is compatible with it. Bob could be smater than average, have a lower optimal sleep length, and have fewer consequences per hour of foregone sleep ). I want to highlight that there are many tasks that don’t benefit from the additional 15 IQ points, and if your total output depends more on those than on tasks that benefit from every last IQ point, then trading IQ points for more hours in the day can be the optimal move. But if Bob got smarter co-workers or became less resilient to sleep loss the trade-off might be different, and it’s beneficial for him to check the trade-offs periodicially.
I don’t think that’s what Alexey or Natalia were arguing, and it ignores health effects, but what I’m excited about in this discussion is rendering these kinds of trade-offs able to reasoned about
I agree about your point on making trade-offs possible to reason about. I’ve found that in discussing this away from LW, people will talk as if extra hours of sleep don’t come at the expense of waking hours. To them, at first glance, the decision is between “activity X, Y, and Z with 6 or 8 hours of sleep” rather than between “activity X and Y with 8 hours of sleep, or activity X, Y, and Z with 6 hours of sleep.”
When I point this out, the follow-up is “but what if the lost sleep makes activity Z a negative experience?” When we eventually anchor on that being the question at hand, rather than the assumption by default, the conversation can become productive and interesting.
I also agree with your bottlenecks-based framing. I think some tasks are bottlenecked by cognitive function (chess competitions), others usually aren’t but can be in rare/emergency circumstances that therefore require full-time high cognitive function (airline pilots), and many others aren’t bottlenecked by cognition (sailing races).
One counterpoint to that framing is that activities usually don’t take a fixed amount of time. There are many things that I’ve done where I’ll take a quarter of the time or less if I’m in a productive state rather than a non-productive state. Sometimes it might really be a trade off between “activity X, Y, and Z with 6 or 8 hours of sleep”, or even “activity X, Y, and Z with 8 hours of sleep, or activity X and Y with 6 hours of sleep”. I’ve also found that I’m less distracted with more sleep, so it could be “be productive for the most of the day with 8 hours of sleep, or browse social media for most of the day with 6 hours of sleep”. This all has been most obvious when I’ve been working on intellectual activities, like writing or developing software, but I also tend to have slower reflexes when I’ve been sleeping less, so physical activities are also relevant.
My personal sleep logs show that productive states are strongly correlated with the amount of hours of sleep I’ve gotten over the last month and the previous night. I don’t have proof that the math really works out, but I strongly suspect it does. Over longer term periods where I’ve had more or fewer hours of sleep, I’ve gotten more positive outside feedback (grades, feedback in monthly reviews) in the the periods with more sleep. But of course I have no proof that this applies to everyone, it’s correlation instead of causation, and it’s certainly no double blind study.
And of course the hardest problem is beating the akrasia to go to bed when I want to, but that’s a separate issue.
Agree! I think we are coming to a model where N hours of SD results in an average of X% loss in productivity per hour, and a change from Y to Z daily productive hours. We want to know the sign of XZ—Y.
Here, you may be arguing both that the performance loss is so great as to more than cancel out the increased number of working hours. You may also be arguing that we actually lose productive hours with sleep deprivation, despite being awake longer, because the loss of executive function makes us unable to focus on our work.
Alternatively, it may be that the additional productive hours from SD can be put into low-demand but time-consuming life maintenance or pleasurable occupations that otherwise might get neglected. For example, we can imagine a busy student choosing between 8 hours of sleep, or 6 hours of sleep plus 2 hours of socializing. Or between 8 hours of sleep, or 7 hours of sleep plus a trip to the grocery store. Or between 8 hours of sleep and 7 hours of sleep plus 1 hour of exercise.
Sacrificing sleep for efficiency-compounding tasks and meaningful occupations may cancel out or even reverse the cognitive impacts of SD itself. It may be that the worst effects of SD can be mitigated with a combination of practice, organizational supports, and stimulating activities and caffeine. It may also be that for some people, their baseline need for sleep is below the level recommended by sleep researchers, so that for them, reduced sleep comes at no cognitive cost.
So the question is twofold. Does non-strategic mild-to-moderate SD lead to net positive or net negative results on some outcome, like productivity, pleasure, health, or meaningful experiences? Furthermore, to what extent can an SD-management strategy mitigate the negative impacts of mild SD?
The stakes are rather significant. Consider that 2 hours of daily SD from age 20-60 adds up to 3.3 years of additional wakefulness, distributed across the most productive and healthy part of the lifespan. If it ruins the experience of life and damages productivity, that is an enormous detriment. On the other hand, if long-term strategically managed SD can be made neutral to positive, then that is the only intervention we currently have to more or less add healthy years to the natural lifespan.
Given the obvious complexities of managing sleep, even for adults people who are attempting to get a full 7-8 hours per night, it seems worth puzzling through it up front in order to establish an optimal sleep routine, given one’s individual preferences, bodily needs, and demands of life.
I tried to be explicit about the average number of hours slept in the studies included in these meta-analyses of experimental studies. Usually, it’s about 4 to 5 hours, with the average seeming to be closer to 4. I also tried to be explicit about how many hours of sleep per night my claims refer to in my conclusion (except for the first claim, which I specified was about “shorter-than-average” sleep, and the last one, which is a harder-to-operationalize claim, but still didn’t explicitly refer to “sleep deprivation”).
Standard deviations are used very often; I would be surprised if they weren’t used in 5% or more of the sleep studies I’ve read.
Sorry yes, of course they’re including SD calculations in their papers. What I meant was that medical papers (or popular coverage thereof) almost never consider the implications of the fact that 0.3% of people will be >=3 standard devs out, and with a large enough population that’s a lot of people.