Is there such a thing as "sound" ethics a la "sound logic"?
December 19, 2021 8:51 PM Subscribe
If not, how close can one get to proving something ethically "true" or not?
Part of the reason I am asking this question is because I want to see what would be necessary for creating a "calculator" which could be used to replace representational government (since itʻs absurd that we are still subjecting ourselves to the whims and opinions of "legislators" like Joe Manchin to make decisions on behalf of humanityʻs future).
This was the idea behind utilitarianism, or really Bentham’s “hedonic calculus”. Given that folks have been at it for 250 years or so (maybe all the way back to the ancient Greeks, in eudemonia) I don’t think it’s really a solvable thing.
posted by supercres at 9:32 PM on December 19, 2021 [2 favorites]
posted by supercres at 9:32 PM on December 19, 2021 [2 favorites]
Best answer: The two primary ethical theories that attempted to present particular ethical claims as having truth values are 1) Utilitarianism, as constructed by Bentham or Mill, probably the closest to a "calculator" in that any given action can be evaluated based entirely on the results-- every claim is easy to evaluate in a rule sense, since it's all based on events, and 2) Deontology, specifically Kantian, which is rooted in necessary logical conclusions based on the existence of beings capable of performing logical reasoning.
Utilitarianism would be better for a "politician replacement" probably, just in terms of computational processes as they currently exist. Programming the calculator to treat rational beings as if their will matters, specifically when they're acting according to respect for other persons, would be a messy proposition since humans tend to argue a lot about what the hell that even means; whereas Utilitarians generally agree on what actions are good, their arguments mostly center around whether it's possible to actually perform good action consistently and how to avoid repulsive conclusions.
posted by Grim Fridge at 9:39 PM on December 19, 2021 [3 favorites]
Utilitarianism would be better for a "politician replacement" probably, just in terms of computational processes as they currently exist. Programming the calculator to treat rational beings as if their will matters, specifically when they're acting according to respect for other persons, would be a messy proposition since humans tend to argue a lot about what the hell that even means; whereas Utilitarians generally agree on what actions are good, their arguments mostly center around whether it's possible to actually perform good action consistently and how to avoid repulsive conclusions.
posted by Grim Fridge at 9:39 PM on December 19, 2021 [3 favorites]
"Soundness" in logic is not exactly the same thing as "truth". A logical argument is sound if it produces a conclusion from a set of premises using only steps that are valid within the logical system. Whether the conclusion of the argument is deemed to be true may depend upon whether one accepts the premises used. For example, the following is a sound argument:
A1. Socrates is a man.
A2. All men are mortal.
A3. Therefore, Socrates is mortal.
However, so is the following:
B1. Socrates is a cheese sandwich.
B2. All cheese sandwiches are mariachi bands.
B3. Therefore, Socrates is a mariachi band.
Both A and B employ the same formal logical structure and are equally sound, but while most people would accept the premises A1 and A2 and therefore the conclusion A3, most people would find B1 and B2 to be absurd premises on which to build an argument, and therefore reject the conclusion B3 as untrue even if it is a sound conclusion given the premises. Furthermore, there are alternative logical systems that actually employ different rules of argumentation, so an argument that is sound in one system may be unsound in another, such that two people who agree on the truth of the initial premises may disagree on whether the conclusion follows if they employ different logical systems.
I am not a philosopher, but as I understand it, much of the history of ethical philosophy in the Western tradition, and especially since the rise of analytic philosophy over the last century or so, has been very concerned with the soundness of ethical arguments in just the same way as it's been concerned with the soundness of other logical arguments. In other words, determining whether an ethical system is internally consistent according to its own rules and premises. The problem with applying this approach at a policy level is that even if everyone were to engage in the process in good faith, not everyone would agree upon which premises are to be accepted (e.g., is a fetus a person from a moral and legal standpoint?) or even which rules to apply (e.g., should we employ consequentialist or deontic reasoning? And what consequences are to be weighed, and how do we measure them? Which deontic rules are to be respected?).
I could opine on how I think the idea of some sort of objective moral calculator as a solution to the problems with American democracy may seem appealing but fundamentally misses the point, but that's not really appropriate for an AskMe answer -- suffice to say that that is my opinion. Logic is a powerful tool for analyzing ethics, but even logical argumentation is always relative to a particular system of rules and axioms, and ethics is even more so. Learning to assess whether an argument is sound with respect to some ethical system is a valuable exercise, but it is only the beginning, not the end, of ethical inquiry.
posted by biogeo at 10:07 PM on December 19, 2021 [19 favorites]
A1. Socrates is a man.
A2. All men are mortal.
A3. Therefore, Socrates is mortal.
However, so is the following:
B1. Socrates is a cheese sandwich.
B2. All cheese sandwiches are mariachi bands.
B3. Therefore, Socrates is a mariachi band.
Both A and B employ the same formal logical structure and are equally sound, but while most people would accept the premises A1 and A2 and therefore the conclusion A3, most people would find B1 and B2 to be absurd premises on which to build an argument, and therefore reject the conclusion B3 as untrue even if it is a sound conclusion given the premises. Furthermore, there are alternative logical systems that actually employ different rules of argumentation, so an argument that is sound in one system may be unsound in another, such that two people who agree on the truth of the initial premises may disagree on whether the conclusion follows if they employ different logical systems.
I am not a philosopher, but as I understand it, much of the history of ethical philosophy in the Western tradition, and especially since the rise of analytic philosophy over the last century or so, has been very concerned with the soundness of ethical arguments in just the same way as it's been concerned with the soundness of other logical arguments. In other words, determining whether an ethical system is internally consistent according to its own rules and premises. The problem with applying this approach at a policy level is that even if everyone were to engage in the process in good faith, not everyone would agree upon which premises are to be accepted (e.g., is a fetus a person from a moral and legal standpoint?) or even which rules to apply (e.g., should we employ consequentialist or deontic reasoning? And what consequences are to be weighed, and how do we measure them? Which deontic rules are to be respected?).
I could opine on how I think the idea of some sort of objective moral calculator as a solution to the problems with American democracy may seem appealing but fundamentally misses the point, but that's not really appropriate for an AskMe answer -- suffice to say that that is my opinion. Logic is a powerful tool for analyzing ethics, but even logical argumentation is always relative to a particular system of rules and axioms, and ethics is even more so. Learning to assess whether an argument is sound with respect to some ethical system is a valuable exercise, but it is only the beginning, not the end, of ethical inquiry.
posted by biogeo at 10:07 PM on December 19, 2021 [19 favorites]
The root of the problem is that there isn't consensus of what moral good actually is. Is the right action the one that causes the most good for the most people? Now you have scenarios where people are being murdered for the greater good (e.g. the trolley problem). Should everyone always act with virtue, being "good" and never lying? Well, now you have someone telling the death squads where the refugees are hiding.
We're good at imagining broad strokes, but things really break down when it gets to specifics. That's why the whole field of ethics exists.
posted by Special Agent Dale Cooper at 10:08 PM on December 19, 2021 [2 favorites]
We're good at imagining broad strokes, but things really break down when it gets to specifics. That's why the whole field of ethics exists.
posted by Special Agent Dale Cooper at 10:08 PM on December 19, 2021 [2 favorites]
"Sound logic" only lets you determine the consequences of axioms. Any "sound ethics" would have the same problem -- someone has to input the numbers into the calculator.
posted by panic at 10:47 PM on December 19, 2021
posted by panic at 10:47 PM on December 19, 2021
Best answer: von Wright’s deontic logic was / is / involves the attempt to use modal operators (box, diamond) to represent obligatoriness and permissibility. The idea overlaps with yours, that by specifying an axiomatised formal system it should be possible to identify which (moral, ethical) arguments or inferences are valid, or to show when counter models exist. You can read more about it here.
posted by Joeruckus at 12:37 AM on December 20, 2021 [1 favorite]
posted by Joeruckus at 12:37 AM on December 20, 2021 [1 favorite]
A logical argument is sound if it produces a conclusion from a set of premises using only steps that are valid within the logical system.
A deductive argument is valid if and only if its conclusion follows from its premises, according to the rules for deduction. It is sound if it is valid AND its premises are true.
posted by thelonius at 2:38 AM on December 20, 2021 [5 favorites]
A deductive argument is valid if and only if its conclusion follows from its premises, according to the rules for deduction. It is sound if it is valid AND its premises are true.
posted by thelonius at 2:38 AM on December 20, 2021 [5 favorites]
AND its premises are true.
Ok so now we have to talk about the various theories of truth…
There aren’t really any shortcuts, unfortunately. I guess that representational democracy as a method of decision making amounts to this: Let’s put all the clever people in a room together & let them work it out, then we’ll do what they say. Problem is that some stupid people also sneak into the room & fuck up the process.
So I feel like what we need is a way to find & remove the stupid (& self-serving & ill-intentioned) people, rather than a moral calculator.
posted by rd45 at 3:42 AM on December 20, 2021 [4 favorites]
Ok so now we have to talk about the various theories of truth…
There aren’t really any shortcuts, unfortunately. I guess that representational democracy as a method of decision making amounts to this: Let’s put all the clever people in a room together & let them work it out, then we’ll do what they say. Problem is that some stupid people also sneak into the room & fuck up the process.
So I feel like what we need is a way to find & remove the stupid (& self-serving & ill-intentioned) people, rather than a moral calculator.
posted by rd45 at 3:42 AM on December 20, 2021 [4 favorites]
I want to see what would be necessary for creating a "calculator" which could be used to replace representational government
It turns out that any such calculator or method (also including representation) is guaranteed to have horrific flaws, and that aggregating people's preferences is a truly intractable problem.
Anything that inputs the preferences of individual people and outputs either a preference statement ("We like A better than B") or a choice ("We choose X over Y") is a social choice mechanism. Here are some things we'd like in a social choice mechanism:
ST -- Social transitivity. If we choose A over B and B over C, we should choose A over C.
U -- Universal admissibility. Nobody is kicked out of participating in the social choice mechanism because of the content of their preferences.
P -- Pareto optimality. If literally everyone prefers F to G, we'd better choose F over G.
I -- Independence from irrelevant alternatives. If an evil wizard inverts everyone's preferences between chocolate and vanilla ice cream, that shouldn't affect the choice between naming your dog Floober and naming it Bitey.
D -- NonDictatorship. There is no single person who determines every single choice our society makes.
And Ken Arrow proved in the 50s that you can't have all 5 of those things at the same time. If you have the first 4, everyone's preferences might line up nicely to prevent an Arrovian dictator, or might line up to create one. ST-U-P-I-D makes a nice mnemonic.
You can certainly do better institutional design than the US did by many metrics, but you should temper your expectations. Even if you could do the work to design a new system of preference aggregation, know in advance that it too is absolutely guaranteed to be terrible.
posted by GCU Sweet and Full of Grace at 6:04 AM on December 20, 2021 [10 favorites]
It turns out that any such calculator or method (also including representation) is guaranteed to have horrific flaws, and that aggregating people's preferences is a truly intractable problem.
Anything that inputs the preferences of individual people and outputs either a preference statement ("We like A better than B") or a choice ("We choose X over Y") is a social choice mechanism. Here are some things we'd like in a social choice mechanism:
ST -- Social transitivity. If we choose A over B and B over C, we should choose A over C.
U -- Universal admissibility. Nobody is kicked out of participating in the social choice mechanism because of the content of their preferences.
P -- Pareto optimality. If literally everyone prefers F to G, we'd better choose F over G.
I -- Independence from irrelevant alternatives. If an evil wizard inverts everyone's preferences between chocolate and vanilla ice cream, that shouldn't affect the choice between naming your dog Floober and naming it Bitey.
D -- NonDictatorship. There is no single person who determines every single choice our society makes.
And Ken Arrow proved in the 50s that you can't have all 5 of those things at the same time. If you have the first 4, everyone's preferences might line up nicely to prevent an Arrovian dictator, or might line up to create one. ST-U-P-I-D makes a nice mnemonic.
You can certainly do better institutional design than the US did by many metrics, but you should temper your expectations. Even if you could do the work to design a new system of preference aggregation, know in advance that it too is absolutely guaranteed to be terrible.
posted by GCU Sweet and Full of Grace at 6:04 AM on December 20, 2021 [10 favorites]
Aside from the theory, there is practice. We know a lot about how an ethical government can be run. The list of best practices is long and includes stuff like sealed bid purchasing and separating accounts receivable from accounts payable and transparency. But most any government body has exceptions, minor ones and major ones like civil asset forfeiture. They exist for reasons of power and greed, and they exist for political reasons like trying to keep programs off-budget to make government spending look smaller than it is.
There is also a problem of unintended consequences. Example: the student loan program was created with good intentions but had the unintended consequence of a generation burdened by debt.
posted by SemiSalt at 9:10 AM on December 20, 2021 [1 favorite]
There is also a problem of unintended consequences. Example: the student loan program was created with good intentions but had the unintended consequence of a generation burdened by debt.
posted by SemiSalt at 9:10 AM on December 20, 2021 [1 favorite]
Also aside from the theory, there's the question of what data the calculator would process. This isn't a "garbage in, garbage out" thing. It's the fact that even when using good quality data, there are still questions around which datasets get used, who decides what data fields to collect and why, and what gets overlooked. I highly recommend reading Data Feminism from MIT Press for a straightforward intro to these issues.
posted by cadge at 9:47 AM on December 20, 2021 [1 favorite]
posted by cadge at 9:47 AM on December 20, 2021 [1 favorite]
what would be necessary for creating a "calculator" which could be used to replace representational government
Well, one of the necessities would be programmers who were both technically and ethically superb. They may exist, but we don’t have a system to identify them.
Also, if people didn’t trust the algorithm they didn’t understand, this would be an awful government. Example: the US already runs an approximate version of this, our legal system, which is an attempt to codify ethics in human-readable language. It has some of the strengths you’re looking for and all the problems above.
posted by clew at 9:57 AM on December 20, 2021
Well, one of the necessities would be programmers who were both technically and ethically superb. They may exist, but we don’t have a system to identify them.
Also, if people didn’t trust the algorithm they didn’t understand, this would be an awful government. Example: the US already runs an approximate version of this, our legal system, which is an attempt to codify ethics in human-readable language. It has some of the strengths you’re looking for and all the problems above.
posted by clew at 9:57 AM on December 20, 2021
Point of order, biogeo's otherwise good explanation is for logical validity, not logical soundness.
Their second example is valid but not sound. Sound logical arguments must have true premises. This is the standard terminology used by logicians and mathematicians and most philosophers.
posted by SaltySalticid at 10:18 AM on December 20, 2021 [4 favorites]
Their second example is valid but not sound. Sound logical arguments must have true premises. This is the standard terminology used by logicians and mathematicians and most philosophers.
posted by SaltySalticid at 10:18 AM on December 20, 2021 [4 favorites]
thelonius and SaltySalticid, thank you for the correction. Anyone who appreciated my response, please take note of their correction.
posted by biogeo at 3:39 PM on December 20, 2021 [1 favorite]
posted by biogeo at 3:39 PM on December 20, 2021 [1 favorite]
Ignoring the technical side of things, when the population can't agree on even very basic 'rules' such as whether it's OK to kill another human, there doesn't seem to be any workable way to implement this. Without clear, agreed-on rules, no such system could work. Without even getting into the issues of complex problems, where multiple rules would have to be applied simultaneously, our inability to create even simple rules would doom this to failure (just like the legal system clew mentions).
An awesome thought experiment, though :-)
posted by dg at 6:26 PM on December 20, 2021
An awesome thought experiment, though :-)
posted by dg at 6:26 PM on December 20, 2021
when the population can't agree on even very basic 'rules' such as whether it's OK to kill another human, there doesn't seem to be any workable way to implement this
Let alone that it can't even agree on what constitutes another human.
Let alone that it contains an enormous proportion of people whose moral calculus extends no further than the last suggestion from a shouting head on Fucks News.
I want to see what would be necessary for creating a "calculator" which could be used to replace representational government
You and Isaac Asimov both.
posted by flabdablet at 6:13 AM on December 21, 2021
Let alone that it can't even agree on what constitutes another human.
Let alone that it contains an enormous proportion of people whose moral calculus extends no further than the last suggestion from a shouting head on Fucks News.
I want to see what would be necessary for creating a "calculator" which could be used to replace representational government
You and Isaac Asimov both.
posted by flabdablet at 6:13 AM on December 21, 2021
This thread is closed to new comments.
posted by tiamat at 9:03 PM on December 19, 2021