I would say that, "Each counts for one," does mean a little more than just anonymity - after all, "Each counts for zero," also satisfies anonymity. So, "Each counts for one," is actually telling us two things: First, that everyone's lives count equally, and second, that everyone's lives count for something. Taurek's view may satisfy the first, but it violates the latter: After the first person in each group whose life is threatened, no one counts for anything if you accept Taurek's view.
In other words, "Each counts for one," is really Anonymity + Positive Responsiveness.
Also, it seems like it was meant less as a proof of number counting and more as a refutation of Taurek's claim that number counting does not reflect an equal concern for each person. Parfit's point is that Taurek is just factually wrong about that - number-counting manifestly does reflect equal concern, even if other views do as well, since it counts each person equally.
I think you’re making a lot of good points. Maybe we can interpret Parfit as saying that “each counts for one” means more than equality—it also includes positive responsiveness.
Though in that case, it’s still true that we can’t use “each counts for one” as a plausible way to argue against Taurek. Positive responsiveness is precisely what’s at issue! To assume it would be to make the argument circular.
Maybe, but like I said, I don't think it was meant as a refutation of Taurek's view, but as a counterargument to Taurek's claim that number-counting views treat people unequally.
If you do want to consider it as a positive argument in favor of number-counting, then you should probably think of it as just pointing out how obvious positive responsiveness is - What kind of ethical theory would deny that each person's life counts for something? Not one we want to endorse.
One problem for Parfit is that he doesn't accept positive responsiveness across the board. He thinks that when A and B are "imprecisely equal" in value, a slight improvement needn't break the tie.
To use an example in the style of On What Matters Vol. 1, suppose you can rush into a burning building to save a life. Parfit thinks the ("impersonal") reasons to rush in are roughly equal to the ("personal") reasons to stay out. But he thinks the same is true even if there's a *second* person in the building you could save.
(This wasn't a problem for him back in 1978 -- only in hindsight, given his later work.)
I agree with you that Parfit's got a good counterargument to Taurek. But in my defense, the ethics lore is that Parfit rekt Taurek, and if you look at the first few footnotes of "Each Counts for One," you'll find prominent applied ethicists saying "As Parfit showed, equality requires us to count the numbers" -- that's almost a direct quote!
I hadn’t seen that meme before—thanks for sharing!
A couple of potential replies;
1) Frances Kamm thinks that some rules would be chosen behind the veil but that, in reality, they don’t hold. (Like “don’t kill one person to redistribute their organs to save five lives.”)
2) Rahul Kumar and Kerah Gordon-Solmon argue that, even behind the veil, there’s more reason than you might think to choose a policy of not killing, because there might be deontological reasons against choosing the policy. (For example, “this policy would be an affront to our humanity.”)
My sense is that most deontologists go for a version of (1). They don’t think veil arguments have much force.
I think you're right but I'd come at it from a slightly different angle. "Each counts for one" -- clear as it may seem -- is ambiguous. Each *what* counts for one? Taurek could easily parry Parfit by saying that each group or cluster counts for one. And since he can save only one group or cluster, considerations of equality give him no more reason to save the one than the other. He'd then be logically committed (I think) to saying that more groups or clusters count for more, which would undermine his position. But I think the real lesson of the Parfit/Taurek debate is that, by itself, a principle of equality doesn't settle questions of individuation. Indeed, it shows that you have to settle on an individuation scheme -- on who or what is to count for one -- before matters of equality can be sensibly discussed.
Hah! I don't know how I'd never thought of it that way. "Each group counts for one" makes a lot of sense.
I do, however, wonder if Taurek would be okay with conceding that some individuals count for more than one, on his view. Maybe he would! But I was assuming he wouldn't.
I think this makes the connection to choice theory even more explicit. One critique of Arrow’s theorem is that it only applies to ordinal preferences/voting. This is, if we *could* make interpersonal utility comparisons as early welfare theorists wanted to, we could adopt some scoring system (rule) that would satisfy Arrow’s conditions (each *util* counts for one). Economists have almost universally conceded cardinal utility for ordinal utility (I’m not yet sure if this is for any more reason than that ordinal utility makes fewer, more plausible assumptions. Maybe these assumptions are *still* unjustified?).
Taking the May-ian formalization of Taurek, we can think of Taurek as going even further than ordinality and stating that we cannot even count up binary-rank-order preferences.
This is what the Positive Responsiveness (monotonicity) assumption is getting at. That adding a vote must not decrease the probability of the outcome. In other words, two people voting for X must outweigh one voting for Y. This is the whole basis of welfare theory with ordinal preferences.
Isn’t this whole argument just about Positive Responsiveness (I guess that’s the lesson here)? Should the marginal vote for X make X more likely to be optimal? The answer is yes if each vote counts for one.
I left two comments on the New Work in Philosophy posting - let me know if I should move them here (I notice that Substack doesn't tend to get discussion).
Incidentally, how do you effectively navigate comments on your own posts in Substack? Do you use the app or the browser? For me, my own posts and comments on my own posts seem to function very differently in-browser from other people's posts and comments on other people's posts. When I click the "view comment" button in an e-mail, it takes me to a page with the first several lines of the comment, and a text box for me to write a reply, but I usually can't see the whole comment, and definitely can't see other replies to the comment.
I usuall find the emails useless except as a reminder that I wrote something recently. Most of my replies are in the app or just from the notifications I get while 'stacking on my laptop.
(Though admittedly I got to *this* comment via email...)
I've twice tried installing the app, because Substack is pushing it, but comment threads on other people's posts are much harder to deal with inside the app (both times I checked). Maybe I'll need one device that has the app, just for dealing with my own comment threads...
I would say that, "Each counts for one," does mean a little more than just anonymity - after all, "Each counts for zero," also satisfies anonymity. So, "Each counts for one," is actually telling us two things: First, that everyone's lives count equally, and second, that everyone's lives count for something. Taurek's view may satisfy the first, but it violates the latter: After the first person in each group whose life is threatened, no one counts for anything if you accept Taurek's view.
In other words, "Each counts for one," is really Anonymity + Positive Responsiveness.
Also, it seems like it was meant less as a proof of number counting and more as a refutation of Taurek's claim that number counting does not reflect an equal concern for each person. Parfit's point is that Taurek is just factually wrong about that - number-counting manifestly does reflect equal concern, even if other views do as well, since it counts each person equally.
I think you’re making a lot of good points. Maybe we can interpret Parfit as saying that “each counts for one” means more than equality—it also includes positive responsiveness.
Though in that case, it’s still true that we can’t use “each counts for one” as a plausible way to argue against Taurek. Positive responsiveness is precisely what’s at issue! To assume it would be to make the argument circular.
Maybe, but like I said, I don't think it was meant as a refutation of Taurek's view, but as a counterargument to Taurek's claim that number-counting views treat people unequally.
If you do want to consider it as a positive argument in favor of number-counting, then you should probably think of it as just pointing out how obvious positive responsiveness is - What kind of ethical theory would deny that each person's life counts for something? Not one we want to endorse.
One problem for Parfit is that he doesn't accept positive responsiveness across the board. He thinks that when A and B are "imprecisely equal" in value, a slight improvement needn't break the tie.
To use an example in the style of On What Matters Vol. 1, suppose you can rush into a burning building to save a life. Parfit thinks the ("impersonal") reasons to rush in are roughly equal to the ("personal") reasons to stay out. But he thinks the same is true even if there's a *second* person in the building you could save.
(This wasn't a problem for him back in 1978 -- only in hindsight, given his later work.)
I agree with you that Parfit's got a good counterargument to Taurek. But in my defense, the ethics lore is that Parfit rekt Taurek, and if you look at the first few footnotes of "Each Counts for One," you'll find prominent applied ethicists saying "As Parfit showed, equality requires us to count the numbers" -- that's almost a direct quote!
Speaking of Trolley Problem Memes, I've yet to see a good counterargument to this meme re: aggregation. https://images.huffingtonpost.com/2016-06-01-1464822100-6800400-Screenshot20160601at16.00.34.png
I think it presents the case very cleanly, simply, and intuitively
I hadn’t seen that meme before—thanks for sharing!
A couple of potential replies;
1) Frances Kamm thinks that some rules would be chosen behind the veil but that, in reality, they don’t hold. (Like “don’t kill one person to redistribute their organs to save five lives.”)
2) Rahul Kumar and Kerah Gordon-Solmon argue that, even behind the veil, there’s more reason than you might think to choose a policy of not killing, because there might be deontological reasons against choosing the policy. (For example, “this policy would be an affront to our humanity.”)
My sense is that most deontologists go for a version of (1). They don’t think veil arguments have much force.
I think you're right but I'd come at it from a slightly different angle. "Each counts for one" -- clear as it may seem -- is ambiguous. Each *what* counts for one? Taurek could easily parry Parfit by saying that each group or cluster counts for one. And since he can save only one group or cluster, considerations of equality give him no more reason to save the one than the other. He'd then be logically committed (I think) to saying that more groups or clusters count for more, which would undermine his position. But I think the real lesson of the Parfit/Taurek debate is that, by itself, a principle of equality doesn't settle questions of individuation. Indeed, it shows that you have to settle on an individuation scheme -- on who or what is to count for one -- before matters of equality can be sensibly discussed.
Hah! I don't know how I'd never thought of it that way. "Each group counts for one" makes a lot of sense.
I do, however, wonder if Taurek would be okay with conceding that some individuals count for more than one, on his view. Maybe he would! But I was assuming he wouldn't.
I think this makes the connection to choice theory even more explicit. One critique of Arrow’s theorem is that it only applies to ordinal preferences/voting. This is, if we *could* make interpersonal utility comparisons as early welfare theorists wanted to, we could adopt some scoring system (rule) that would satisfy Arrow’s conditions (each *util* counts for one). Economists have almost universally conceded cardinal utility for ordinal utility (I’m not yet sure if this is for any more reason than that ordinal utility makes fewer, more plausible assumptions. Maybe these assumptions are *still* unjustified?).
Taking the May-ian formalization of Taurek, we can think of Taurek as going even further than ordinality and stating that we cannot even count up binary-rank-order preferences.
This is what the Positive Responsiveness (monotonicity) assumption is getting at. That adding a vote must not decrease the probability of the outcome. In other words, two people voting for X must outweigh one voting for Y. This is the whole basis of welfare theory with ordinal preferences.
Isn’t this whole argument just about Positive Responsiveness (I guess that’s the lesson here)? Should the marginal vote for X make X more likely to be optimal? The answer is yes if each vote counts for one.
I left two comments on the New Work in Philosophy posting - let me know if I should move them here (I notice that Substack doesn't tend to get discussion).
Sorry, I didn't see those because I don't get notifications for it! I'll respond over there :)
Incidentally, how do you effectively navigate comments on your own posts in Substack? Do you use the app or the browser? For me, my own posts and comments on my own posts seem to function very differently in-browser from other people's posts and comments on other people's posts. When I click the "view comment" button in an e-mail, it takes me to a page with the first several lines of the comment, and a text box for me to write a reply, but I usually can't see the whole comment, and definitely can't see other replies to the comment.
I usuall find the emails useless except as a reminder that I wrote something recently. Most of my replies are in the app or just from the notifications I get while 'stacking on my laptop.
(Though admittedly I got to *this* comment via email...)
I've twice tried installing the app, because Substack is pushing it, but comment threads on other people's posts are much harder to deal with inside the app (both times I checked). Maybe I'll need one device that has the app, just for dealing with my own comment threads...
I’d consider rewriting this, something along the lines of “I love Parfit but he got one thin wrong”.
The biggest part of your article was kind of buried imo