We at PEA Soup are happy to announce another timely entry in The Pebble of broad interest-a short piece on traveling during the pandemic brought to you by Benjamin Hause, a graduate student in philosophy at Florida State University. Many thanks to Benjamin for choosing PEA Soup! Let’s turn it over to Ben:
Last month we saw many condemnations of Thanksgiving travelers. David Faris, of The Week, called them “evil, selfish, and cartoonishly destructive”. And New York Times columnist Farhad Manjoo was the subject of scorn after confessing his Thanksgiving travel plans. Manjoo himself admitted his decision to travel was “like a gut call, ruled more by emotion than empirical data”. Now, just before Christmas, Covid-19 deaths are even higher. Yet, people are still traveling. Is that okay?
I think so. And I don’t think this is just some heedless “gut call” either. To be clear, I don’t condone systemically risky behavior, nor am I against well executed lockdowns. But I do think that occasional travel, like for Christmas, is within the bounds of risk we have always accepted. To illustrate, let’s consider the 2017-2018 flu season in the United States.
In that season, an estimated 61,000 people died from the flu. Half the recorded deaths occurred during just five weeks. That would extrapolate to an estimated 900 daily flu deaths in those weeks. In comparison, there were about 1500 daily recorded Covid-19 deaths before Thanksgiving, 2100 during the peak in April, and 2500 right now. (These are 7-day averages).
Some people think comparing flu estimates with Covid-19 counts is misleading. But we can also look at excess mortality data. Last April, there were 3260 daily excess deaths. During the worst week of the flu endemic, there were 1114 daily excess deaths. Add back in the non-excess flu deaths, and we are again looking at a considerable flu toll.
As you can see, Covid-19 kills two to three times as many people per day as the flu did during its worst weeks. Were it not for mitigation measures, Covid-19 would kill even more. If we assume that people are interacting ten times less due to those mitigation measures, then Covid-19 is causing twenty or thirty times more deaths per interaction than the flu ever caused. By going out during the pandemic, you are thus twenty or thirty times more likely to cause someone’s death compared to prior flu seasons.
That sounds like a big increase. But notice that it is an increase over a number which never caused us any trepidation in the first place. Indeed, someone with twenty times more social interaction than me is also twenty times more likely to cause a flu death. But we don’t let that bother us, because we don’t normally worry about flu risk. Similarly, someone who occasionally goes out during the pandemic poses no more risk to others than someone who parties all the time during flu season. Condemning the former, but not the latter, is simply inconsistent. That’s data, not gut!
Now, my suggestion is that we condemn neither. Frequent socialization during flu season is okay, as is infrequent travel during the pandemic. But that doesn’t excuse our complete negligence of flu risk before. The pandemic has brought to our attention an underappreciated fact: that we can save thousands of lives every year with low cost mitigation measures like mask wearing and flu vaccination. After we finally move past the pandemic, we may continue to save lives year after year.
The next to the last paragraph looks like a Sorites Paradox that needs more support. There’s some small length of my finger (x) I wouldn’t care if it cut off. There’s some multiple (y) of that length such that I would very much mind (xy) of my finger being cut off!
So maybe risk x of partying during flu season is morally permissible, but 20x is not morally permissible. It doesn’t seem a stretch that increasing a risk of killing someone by an order of magnitude might make an action morally impermissible.
My previous comment misunderstood what was going on. I apologize. The phrase that tripped me up was “But notice that it is an increase over a number which never caused us any trepidation in the first place.”
Now I see the claim is that a 20-fold increase of flu death didn’t bother us in holiday seasons in the past, so a behavior that increases risk 20-fold is not something we’ve previously found impermissible.
Thanks for sharing! I’m wondering if the overburdening of hospitals is a reasonable factor to consider here. I agree that we pay little mind to the increased risk of flu deaths when we socialize. However, even a bad flu year like 2017-2018 didn’t increase the rate of hospitalization to the point it was overrun or decreasing the level of care patients could get. There wasn’t even the threat of that happening.
The increased risk brought on by holiday travel thid year isn’t just an increased risk of killing someone, it’s an increased risk of contributing to the dysfunction of what we could consider necessary infrastructure. If that makes for a morally significant difference between the two, then it could still be plausible that traveling during flu season is okay but traveling during the pandemic is not.
In addition to the other criticisms I just want to point out the silliness in resisting changes in present behavior because of past behavior. There is no sense in resisting the influence of new data and new principles, as why we once had no trepidation may have merely been ignorance or incompetence. Seeking consistency with past ignorant moral opinions is irrational, and overmines the value of consistency with past precedents.
As the saying goes, one philosopher’s modus ponens is another’s modus tollens.
Ben writes, “Covid-19 is causing twenty or thirty times more deaths per interaction than the flu ever caused.” He reasons that since flu deaths don’t both us, neither should Covid-19.
But maybe both should. I, for one, didn’t realize pre-pandemic just how deadly the flu season is each year. I’m inclined to take greater precautions in the future during flu season so as not to contribute to those deaths. Come to think of it, I already always get a flu vaccine each year in order to mitigate those effects. But, of course, we won’t yet have any COVID-19 vaccine readily and widely available before holiday travel this year.
On the Utilitarian front I can’t see why we are only looking at deaths and not reduced quality of life impacts.
The post did get me interested in the premise that if people are ok with a doing something that has an x% chance of causing death then they should be, or at least can be predicted to be, ok with an activity that has a 30x% chance of causing death.
Did the author look into the risks of drunk driving vs. sober driving causing death? What about other relevant cases (wearing seatbelt vs not, and non-driving analogies)? I mean if the author is committed to saying that it is ok for us to all drive drunk this holiday season because we are already ok with people driving long distances, that seems like a reductio to me.
A reminder to all about Soup’s comments policy: Full (real) names required from commenters, or the comment gets deleted. Thanks for complying!
Coming back to Scott Looney’s comments–I guess the 20x argument he finds in the post still feels like a Sorites to me, or in any case to be unconvincing. Consider this argument, which seems plainly unconvincing:
Premise 1: One grain of sand is not a beach.
Premise 2: 20 times one grain of sand is still not a beach.
Conclusion: Multiplying the number of grains of sand by 20 from any number cannot make something a beach.
@ Looney, @ Sobel, I don’t think the author wants to make the argument that Looney suggests, although I can see how it might look that way. The only thing that really matters to Hause, I think, is this: “Similarly, someone who occasionally goes out during the pandemic poses no more risk to others than someone who parties all the time during flu season.” If the two behaviors are equally dangerous, then there’s no good reason to condemn the one but not the other. We do not condemn the one, therefore we ought not condemn the other. At least two elements in the argument seem worthy of challenge. As Kohls and Cokelet point out, there are reasons to believe that the two behaviors in question are not equally dangerous (and also reasons not to believe that the two are equally dangerous). And as Ferri and King point out, the right lesson to take away might be that we should condemn highly risky behavior during a flu season.
Hi all, thanks for the excellent comments! I’ll have to think about some of them more, but here are my initial reactions.
@Stacy Kohls: I think that hospital burden is definitely relevant! It is relevant insofar as when you go out, if the hospitals are overburdened, your individual action is more likely to cause harm. This increased risk, if the hospitals are overburdened, should be reflected in the daily death rate, though the daily death rate will probably lag behind hospital burden by some weeks.
@Spencer Ferri, @Matt King: I agree that our past apathy towards the flu was unjustified! But I think that is consistent with my argument. Perhaps the restatement of my argument below, in reply to Scott Looney and David Sobel will be helpful. I take it that you guys would reject premise (1).
@Brad Cokelet: That sounds like a reductio to me too, but I think I can avoid that commitment. I think something like the following principle is true:
You shouldn’t engage in risky behavior unless you have good reason to.
My post focused on showing that occasional travel during the pandemic is no more risky than other behavior that we also do for social reasons. But in your reductio, the reasons at play aren’t of similar quality: there aren’t very good reasons to drive drunk, but there are good reasons to drive long distances (e.g. to visit family).
Returning to my original argument, quarantiners in fact have better reason to engage in their risky behavior (traveling for Christmas) than a frequent flu partier. Social isolation can be depressing, and Christmas is an important tradition for many families. And so mental well-being and family tradition are at sake. In contrast, someone who parties every night during flu season doesn’t seem to miss out on much by partying a bit less. Here, only some marginal entertainment is at sake.
@Scott Looney, @David Sobel: I don’t think I am making the sorites argument. The sorites argument would be like this:
1. We don’t care at all about X flu risk.
2. So we also shouldn’t care about 20X flu risk.
3. Covid risk is 20x flu risk.
4. So we shouldn’t care about Covid risk.
And you both correctly point out that the inference from 1 to 2 is invalid. But my argument is (crudely) more like this:
1. We’d rather put up with X flu risk than believe we ought to regularly limit the number of our social interactions during flu season.
2. Occasional travel during the pandemic also has X risk.
3. So consistency requires that we put up with X Covid risk rather than believe we ought not travel at all.
Now that is technically also invalid for the reasons Brad Cokelet mentioned, but I think (hope) my response there is plausible!
@Olivia Bailey: I saw your response after I typed mine, but I think that is an excellent summary of the dialectic
Sorry, forgot to attach my name to my previous comment!
(@Olivia Bailey: I saw your response after I typed mine, but I think that is an excellent summary of the dialectic)
Let me just follow up on this: “Returning to my original argument, quarantiners in fact have better reason to engage in their risky behavior (traveling for Christmas) than a frequent flu partier. Social isolation can be depressing, and Christmas is an important tradition for many families. And so mental well-being and family tradition are at sake. In contrast, someone who parties every night during flu season doesn’t seem to miss out on much by partying a bit less. Here, only some marginal entertainment is at sake.” Will the relevant contrast be between “partying every night” and “partying a bit less”? If “partying a bit less” is all that it would take to get a dramatic reduction in flu deaths, then it seems like we should *absolutely* condemn the every-night partier– they can easily adjust their behavior to be slightly less social. We’d also then restore consistency without reversing our condemnation of the covid traveller. But if in fact what the partier is required to give up is the whole social lifestyle, well then, I’m not convinced that your intuition that traveling home is responsive to much better reasons is all that attractive. College students absolutely suffer from the lack of social bonding with their peers- it does have a significant impact on well-being.
@Olivia Bailey Thanks again! You’ve convinced me that we should condemn some flu partying if the numbers work out that way. But whether that restores consistency with condemning Christmas travel depends on what amount of flu partying is permissible.
For example, suppose a ’17-’18 flu partier was deciding between two options:
1. Party every other night during flu season, imposing risk X.
2. Party every night during flu season, imposing risk ~2X.
If outings are twenty times more dangerous during the pandemic, the pandemic partier has these similar risk options:
1*. Party once every forty nights, imposing risk X
2*. Party once every twenty nights, imposing risk ~2X.
Suppose we should condemn (2), but not (1). Even if we condemn (2), it doesn’t follow that we should condemn (2*), since the pandemic partier has better reason to go out, since they go out less often. But suppose that is not an important difference, and we should condemn (2*) too. We should still not condemn (1*), if we condone (1). And (1*) is enough to permit Christmas travel, even if someone already did some traveling/partying before.
Also, I think I was mistaken to assume that a certain behavior (e.g. party-going) is automatically twenty to thirty times more dangerous during the pandemic. That is only the case if all other behavior is equal. But if our pandemic partier is quarantining between parties, and the flu partier is not, other behavior is not equal. Quarantining between parties will lower the probability that your party interactions themselves cause harm. So X and ~2X in 1* and 2* are overestimates, perhaps dramatically so.
Thanks for the response and clarification!
I am not clear on why you think the relevant hospital impacts will show up in the death numbers. I know people who want surgery to deal with serious painful and harmful conditions who can’t get those surgeries or who have extra ordinary wait times. And then there are the burdens on the hospital workers working. Shouldn’t these and other negative effects (which do not occur when people party during flu season) be counted?
Thanks for an interesting post and discussion!
@BradCokelet asks in an earlier comment why we wouldn’t also consider reduced quality of life when making a utilitarian assessment of the options, and I think this is a really important relevant difference between calculations about Covid and the flu.
From what I understand (and someone please correct me if I’m wrong!), even very bad seasonal flus don’t generally lead to months-long (and potentially even lifelong) chronic illness in people who have cases of flu that don’t lead to death. And there are lots of reports of so-called Covid “long haulers”: people who remain ill for months and months, or who have serious organ damage, or neurological effects.
We are reasoning about Covid risk under a huge amount of uncertainty about what the medium and long-effects of the virus are, and I worry that this is under appreciated in a lot of news coverage and political talk about the virus and our responses to it. The options with Covid aren’t just “X% of people die” and “Y% fully recover”: there’s some unknown (potentially pretty significant) percentage of people who get Covid whose health will never fully recover.
I suspect that if flu were similar – if 10% or 20% of the people who contracted mild or moderate cases of flu were left with intense chronic fatigue, or ongoing GI symptoms, or newly diagnosed anxiety/depression, etc. for months or even years, our reasoning about balancing risks and benefits with regard to flu season would have to be really different.
@Brad Cokelet @Alida Liberman Thanks for pushing this issue, I worry about it too. While it is easy to quantify deaths, it is harder to quantify impacts on well-being. If the expected harm to people’s well-being of going out during the pandemic is not more than 20-30 times higher than the expected harm to people’s well-being of going out during flu season, then I think my argument still goes through. (X from my previous comment would represent total expected harm, not just expected deaths.) But if the expected harm is higher than that, then my argument does underestimate Covid risk in a way that would weaken my argument. Since that is a real possibility, and since there are other uncertainties, I agree that we should be careful not to conclude too much. But I don’t think I am, in concluding that Christmas travel may be part of a pattern of permissible behavior. I’d be less confident in concluding that traveling every twenty or forty days is permissible (per 1* or 2*), for the aforementioned reasons.
I know people who want surgery to deal with serious painful and harmful conditions who can’t get those surgeries or who have extra ordinary wait times. And then there are the burdens on the hospital workers working. That is only the case if all other behavior is equal. But if our pandemic partier is quarantining between parties, and the flu partier is not, other behavior is not equal. Quarantining between parties will lower the probability that your party interactions themselves cause harm. ornaments