Can an objective metric for good and bad exist?

[Hedonic Treader]

The idea behind this is that good and bad are valuations within the mind, and therefore subjective. However, the existence and structure of minds are objective facts about the physical world, and such facts could be analyzed by a hypothetical metric we could call objective if it were precise and reliable enough.

For instance, when a person is in strong pain, or experiences strong pleasure, or strong repulsion by another's act which it observes as immoral, most scientists and philosophers would agree that there's something objectively measurable happening in that person's brain, and it could be identified, described, and maybe even quantified by an objective analytical metric.

Following this premise, it raises the question if there's a reliable way of deriving such a metric that's applicable to the human brain, and if it could be abstracted and generalized systematically in such a way that it could ultimately be applied as an objective descriptive tool to all kinds of physical systems, in order to determine the existence and maybe even quantity of good or bad within the scope of these systems.

Why do brains feel bad (or good)? A short answer is because these feelings evolved to motivate generally adaptive behavoir. A deeper answer would be to look at how brains feel bad (or good). We could describe exactly what's happening in the brain when it feels strongly bad (or good) and then, as a second step, generalize the description of that phenomenon so that we would recognize the hallmarks of its valuation-bearing informational properties in all other potential physical systems of equal complexity. Since ultimately, any idea of good and bad in human culture stems from how human brains feel about things, a generalizable descriptive metric of that phenomenon could be the ultimate objective description tool for good and bad in the universe.

In relation to this idea, I would like to pose these questions to discuss:

1) Is such a metric possible conceptually?
2) What steps could be taken to derive such a metric in practice?
3) If successful, would having such a metric close the is-ought-gap in ethics?

[Vaniver]

To go from subjective to objective you need uniformity. Objective physical laws are uniform in their application.

So even if you could go inside my brain and measure my utility function, would that give you the uniform human utility function? No, you would need to measure many human brains.

And it seems that if we all agreed about the "ought" on some objective level, we would also all agree about the "ought" on a subjective level. The laws individual particles follow are the same as laws general particles follow. If an individual human is different from humans in general, that subjectivity cannot be removed without removing that difference.

[Dark567]

1) Is such a metric possible conceptually?
2) What steps could be taken to derive such a metric in practice?
3) If successful, would having such a metric close the is-ought-gap in ethics?

1. Is such a metric possible conceptually? Most philosophers think so. I am going to use the common example of Utilitarianism. It believes that pleasure is something that is objective, and that it is 'good'(even though what causes that pleasure could be subjective i.e. I like cookies and you don't). Your example, as I understand, would also be including the effects on a persons brain from the knowledge that an act happened.

2. I don't know.

3. No. Even if we found a universal objective fact about how all people rationalize morality, we have still only described the situation. Why would everyones brain objectively feel good, actually mean we ought to do that thing? We still have to make the assumption that everyones common brain phenomena ought to be done. It seems like a very intuitive assumption, but it is actually still an assumption with just as little evidence as most other moral theories. The is-ought gap is pretty hard to bridge.

Since ultimately, any idea of good and bad in human culture stems from how human brains feel about things

This is not a uncontroversial statement. Philosophers are pretty divided into two camps on this. What you seem to be claiming is called Moral Anti-Realism, which is in contrast to Moral Realism, which holds that there are mind-independent moral facts. If you want to make this an assumption thats find, you wouldn't be alone, but just realize there is considerable debate about whether or not it is true.

[Charlie!]

And it seems that if we all agreed about the "ought" on some objective level, we would also all agree about the "ought" on a subjective level. The laws individual particles follow are the same as laws general particles follow. If an individual human is different from humans in general, that subjectivity cannot be removed without removing that difference.

Well, there's the analogy to the laws of thermodynamics, which are the "best" laws for large numbers of particles, but can't be applied to lone particles.

I would agree though that just averaging peoples brains shouldn't give you rules that are somehow special. They'd have special properties like minimizing the amount people disagreed with them, but to claim that they're "the best" requires additional assumptions.


A more practical problem might be that in order to fully assess the consequences of an action (e.g. killing one person to save two) by measuring peoples' brains, you need to spend an infinite amount of time, AND somehow manage to measure the counterfactual. Real-world measurement can still only give us approximations as to what's right and wrong, and hey, we can do that already.

[Hedonic Treader]

To go from subjective to objective you need uniformity. Objective physical laws are uniform in their application.

So even if you could go inside my brain and measure my utility function, would that give you the uniform human utility function? No, you would need to measure many human brains.

Which we can do. I didn't expect that you can derive a generalizable metric by looking at just one instance of the class of entities in question.

If an individual human is different from humans in general, that subjectivity cannot be removed without removing that difference.

Yes... but given sufficient information, an adequately powerful descriptive metric should be able to account for or even predict that difference. For instance, the metric should be able to tell us (given all the relevant data), that the average person A will consider strong nociceptive sensory input a bad thing, whereas pain asymbolia patient B will not. Similarly, it should be able to tell us that compassionate person X will react negatively to mirror neuron stimuli caused by the suffering of a naked woman being tortured, while sexual sadist Y may react positively to it instead, and furthermore, it should be able to predict that X will feel that Y's reaction itself is a bad thing when hearing about it, and so on.

The key here is that all these facts should be deducible by applying the generalized metric to a sufficiently rich set of data about the specific nature of the individuals in question, treating their individual brains' informational properties as specific instances of the total class of complex informational valuation systems. And if it's abstract enough, it should ideally be applicable to non-human animal brains (and their valuations), and very ambitiously, even alien or artificial minds or very abstract non-personal systems, as long as they contain the general information-theoretic properties of valuation (a continuum of good and bad).

1. Is such a metric possible conceptually? Most philosophers think so. I am going to use the common example of Utilitarianism. It believes that pleasure is something that is objective, and that it is 'good'(even though what causes that pleasure could be subjective i.e. I like cookies and you don't). Your example, as I understand, would also be including the effects on a persons brain from the knowledge that an act happened.

Yes, and more generally, any information, or sensory input, or internal state, that would be perceived as good or bad.

3. No. Even if we found a universal objective fact about how all people rationalize morality, we have still only described the situation. Why would everyones brain objectively feel good, actually mean we ought to do that thing?

Ok, let me give this my best shot. Imagine you're in strong pain, significantly more severe than you would normally be able to accept voluntarily, and you could simply push a button to make it stop and restore you to a pain-free state. Would there be any doubt for you that you "ought" to push that button immediately? I don't think so; the very nature of the badness of the pain would compel you to. Similary, if you're in strong pain, and another agent Z could push the button, but you couldn't, would there be any doubt for you that Z "ought" to push it right away? I think the answer is no. I think you would have no other choice but to see it this way.

Of course, Z could disagree. But now imagine he would have all the necessary information about your situation and your brain's nature, and he would have a descriptive metric that would inform him you're in a "bad" state but only until he pushes the button (without negative consequences for anyone else). Assume here that Z is not compassionate in the traditional sense, ie. he does not feel bad from the mirror neuron inputs by perceiving your suffering, all he has is the knowledge about your nature and the situation, informed by the descriptive metric I hypothezised. Since the (conditional) bad state in this scenario is an objective fact about the world, and Z knows this because he's adequately informed by the metric, would it be logically flawed by Z to deduce he "ought" to push the button?

[Vaniver]

I didn't expect that you can derive a generalizable metric by looking at just one instance of the class of entities in question.

I left unspoken the assumption that different people have different utility functions.

Objective means "independent of observer". Subjective means "dependent on observer." If utility functions depend on the agent calculating it, then we cannot have an objective utility function.

If a large group of third parties could objectively measure my utility function, what value would that add? People would know what I preferred. But I could have solved that problem more simply by preferring those things myself.

Since the (conditional) bad state in this scenario is an objective fact about the world, and Z knows this because he's adequately informed by the metric, would it be logically flawed by Z to deduce he "ought" to push the button?

Yes. Hume's argument (which I imagine is the relevant is-ought gap) is that in order to state "Agent Z should perform action A" you do not just need to know the results of action A- you need to have some goal in mind. "Agent Z should perform action A because it will reduce suffering." Knowing more about action A's consequences won't help you prove that Z should care about reducing suffering.

[Hedonic Treader]

Objective means "independent of observer". Subjective means "dependent on observer." If utility functions depend on the agent calculating it, then we cannot have an objective utility function.

Well, the idea of an objective utility function refers to the 3rd question only. Even if assume that we cannot have an objective utility function, this doesn't rule out the possibility of an objective metric for subjective utility functions, as the nature and existence of subjects with such utility functions are objective facts about the world.

If a large group of third parties could objectively measure my utility function, what value would that add?

If by "value" you mean objective value, you have implicitly answered question 3 in the affermative, which I don't think is your position. If you mean subjective value, than the answer would entirely depend on the utility function of those agents who put the metric to use. If they're universal sadists, they may want to cause the exact opposite of your utility function, if they're preference utilitarians, they would probably prefer if your preferences were met (in a universally well-aligned utility function, even if they can't logically deduce it from objective facts).

People would know what I preferred. But I could have solved that problem more simply by preferring those things myself.

What problem? To meet your preferences? Unless you answer question 3 in the affermative, that's not necessarily the problem people want to solve; the problem I raised in the other questions was exclusively concerned with the feasibilty of a generalized metric that one could use to learn what you preferred, given sufficient objective data about your physical structure, not necessarily any specific application of that metric.

Since the (conditional) bad state in this scenario is an objective fact about the world, and Z knows this because he's adequately informed by the metric, would it be logically flawed by Z to deduce he "ought" to push the button?

Yes. Hume's argument (which I imagine is the relevant is-ought gap) is that in order to state "Agent Z should perform action A" you do not just need to know the results of action A- you need to have some goal in mind. "Agent Z should perform action A because it will reduce suffering." Knowing more about action A's consequences won't help you prove that Z should care about reducing suffering.

Ok, I concede that it doesn't logically bridge the gap. However, it is still relevant to highlight that an observer moment which contains suffering (or the "bad" generally) by definition contains the valuation that it "ought not" to be in that state. If this is true (and I think it is), then the is-ought-gap is indeed closed for all observer moments which contain (a certain degree of) the bad: From the perspective of such an observer moment, the valuation that this state "ought" to be ended is a hard and unambigous ontological fact. An observer moment containing suffering cannot not care about it.

[Vaniver]

Ok, I concede that it doesn't logically bridge the gap. However, it is still relevant to highlight that an observer moment which contains suffering (or the "bad" generally) by definition contains the valuation that it "ought not" to be in that state. If this is true (and I think it is), then the is-ought-gap is indeed closed for all observer moments which contain (a certain degree of) the bad: From the perspective of such an observer moment, the valuation that this state "ought" to be ended is a hard and unambigous ontological fact. An observer moment containing suffering cannot not care about it.

You still haven't bridged that gap.

(This is why, by the way, a lot of people don't like Hume. This sort of thing is better accepted despite being logically unsupported and moving on.)

Let us imagine a machine that knows exactly the state of anyone's mind, and spits out a letter that is associated with that state. All this tells you if that there are two people, and the machine spits out, say, "A" for both of them, they have the same brain state, and if the machine spits out, say, "A" for one and "B" for the other they have different brain states. The machine is able to normalize for each person's subjective preferences, such that if any person prefers "A" to "B" all people will prefer "A" to "B".

Note that that last sentence was the only one to use the word "prefer," and even there I jumped the gun. We don't have any reason to compare brain states by preference, yet. It's only when we start looking at states and saying "ok, people in state "E" exhibit a number of positive behaviors, and when I am in state "E" I feel extremely happy" that we bring in the concept of preferences. And even when we know that "E" is a happier state than "D", that still does not give us a reason to say that people ought to be in state "E" instead of state "D", or that we ought to put others in state "E" instead of state "D" if possible at no cost.

Those "ought" statements only arise from normative statements about goals, not descriptive statements about reality. The goal "I want to be as happy as possible" is a subjective goal- one can generalize it to "I want everyone to be as happy as possible" but generalizing a subjective goal does not make it an objective goal. Your "hard ontological fact" comes from your label of the brain state, not your correct identification of it, and thus the is-ought gap between descriptions and labels remains unbridged.*

*It is at this point that I would set down Hume, pick up Nietzsche, read "4. The falseness of an opinion is not for us any objection to it: it is here, perhaps, that our new language sounds most strangely. The question is, how far an opinion is life-furthering, life-preserving, species-preserving, perhaps species-rearing, and we are fundamentally inclined to maintain that the falsest opinions (to which the synthetic judgments a priori belong), are the most indispensable to us, that without a recognition of logical fictions, without a comparison of reality with the purely imagined world of the absolute and immutable, without a constant counterfeiting of the world by means of numbers, man could not live- that the renunciation of false opinions would be a renunciation of life, a negation of life. To recognize untruth as a condition of life; that is certainly to impugn the traditional ideas of value in a dangerous manner, and a philosophy which ventures to do so, has thereby alone placed itself beyond good and evil." and get on with my life.

[guenther]

First, if we say "Goodness is as described in the Bible", isn't this an objective description of morality (with the exception of gray areas of interpretation)? We can all independently measure what the Bible says. Of course this isn't derived from natural phenomenon like gravity is.

Second, I do think that our morality follows some natural rules that are represented in the state of all humanity (not simply within one brain). It's the same way that languages have to follow some basic rules to have any value (i.e. common words shouldn't be 50 syllables long). I think this is measurable, but it would be a descriptive notion of morality, not prescriptive.

I don't think oughtness is well defined as a natural process. Even if we discovered the creator of the universe and he said "Do X", one could still say "Well that's just his opinion of what I should do". Our notion of what we ought to do comes from us submitting to a body of wisdom.

[morriswalters]

Why do brains feel bad (or good)? A short answer is because these feelings evolved to motivate generally adaptive behavoir. A deeper answer would be to look at how brains feel bad (or good). We could describe exactly what's happening in the brain when it feels strongly bad (or good) and then, as a second step, generalize the description of that phenomenon so that we would recognize the hallmarks of its valuation-bearing informational properties in all other potential physical systems of equal complexity. Since ultimately, any idea of good and bad in human culture stems from how human brains feel about things, a generalizable descriptive metric of that phenomenon could be the ultimate objective description tool for good and bad in the universe.

Look at Saints and Serial Killers two endpoints on the curve of empathy. Suppose you came up with a equation that could accurately describe the states between the two endpoints, what would that give you? Where does "good" fall on that span? Having said that, it should be possible to, at least in a broad sense, to describe the biological mechanism which creates the emotional response. Find and describe the succinct biological responses that make up all emotional states. At least some of the basics are known. Reflexive responses, hormonal triggers are two that come to mind. I assume there are a finite number of these mechanisms. Once that is done it should be possible to define a matrix of basic responses that when mixed and matched make up the basic emotions. The response would vary in intensity across the broad population but the basic triggers would be the same assuming a biological basis that humans share in common. Once that was done one metric would be to define good and bad as peaks in the norms of the curve assuming a normal distribution.

I have no real idea of how you might do this. But it seems that you would have to define all of the basic emotional states. States that I'll call primitive states. These would be states that that are uncomplicated and discreet, if they exist. You would then measure the biological responses to emotions. For instance measure the triggers on new parents when they look at their child. Do this tens of thousands of times. Then analyze the data. Make sure to get the extremes, those who badly want children and those who don't. This should present you with the emotional response of a parent to their child. This should indicate a set of triggers for that emotional response. By adding data over time you might eventually be able to map a normal response for that emotion. Rinse and repeat with other responses you wish to study.

I have no idea if this would give you what you want. But it would mark the range. And you certainly could argue that the undesirable end of the range is bad.

[Hedonic Treader]

Those "ought" statements only arise from normative statements about goals, not descriptive statements about reality. The goal "I want to be as happy as possible" is a subjective goal- one can generalize it to "I want everyone to be as happy as possible" but generalizing a subjective goal does not make it an objective goal. Your "hard ontological fact" comes from your label of the brain state, not your correct identification of it, and thus the is-ought gap between descriptions and labels remains unbridged.

Here's the problem: Descriptive statements about reality don't necessarily imply neutrality. It is the very definition of the good and the bad that they are not neutral. Insofar as it is possible to identifiy them objectively as physical phenomena in the world (such as informational structures within the brain), then we can objectively identify non-neutral ontologically real facts about the world. If something can be objectively true, but without being neutral, it doesn't seem that much of a category mistake to use it as a foundation to derive normative statements.

When you actually experience the good and the bad, you know they're not neutral, and you know they're ontologically real. I disagree that the ontological reality of, say, the badness of strong pain comes from semantic labelling; it is at the very heart of its informational nature (within the context of a system like the brain, which is an objective fact about the world that we can identify correctly within a descriptive method). Or did I misrepresent your point here?

It's only when we start looking at states and saying "ok, people in state "E" exhibit a number of positive behaviors, and when I am in state "E" I feel extremely happy" that we bring in the concept of preferences. And even when we know that "E" is a happier state than "D", that still does not give us a reason to say that people ought to be in state "E" instead of state "D", or that we ought to put others in state "E" instead of state "D" if possible at no cost.

By observing behavior and associated communication acts about internal states, you could build a metric to rank states according to their respective goodness and badness, with at least some degree of precision. You could correlate that with an information-theoretic outline of the informational workings within the brain, ie. you don't just compare static states, but conceptualize the informational algorithms that implement the good and the bad. If this can be done reliably, you have an abstract outline of what it means for a physical configuration to entail goodness or badness as ontologically real properties. And this would not give us reason to say that people (or such systems generally) ought to be brought into the states of higher goodness if possible with no cost?

[Vaniver]

Insofar as it is possible to identifiy them objectively as physical phenomena in the world (such as informational structures within the brain), then we can objectively identify non-neutral ontologically real facts about the world.

This is the sticking point, I think. It is not clear to me that you can identify a brain state with "good" or "bad" without bringing a subjective perspective on board. That would suggest it's not objective.

Now, you might be happy with a definition of objectivity that starts with a subjective position and then builds outwards- but note that in this scenario it requires that we have a magic machine whose correctness we established a priori. If all I knew was how my brain states felt, how the machine labeled my brain states, other people's descriptions of how they feel, and how the machine labeled their brain states, would that be enough to formally know that it applied to every and was correct for everyone? It's good enough for real life- but that's a different thing.

[SnakesNDMartyrs]

An objective metric for good and bad can't exist because nothing is inherently good or bad.

An objective metric for subjective valuations of good and bad could, in theory, exist; although it would have to be continually re-calibrated as the subjective valuations change with time.

[guenther]

An objective metric for good and bad can't exist because nothing is inherently good or bad.

Economic value doesn't inherently exist, but we do have an objective measure for the economic value of stocks, commodities, etc.

[lutzj]

An objective metric for good and bad can't exist because nothing is inherently good or bad.

Economic value doesn't inherently exist, but we do have an objective measure for the economic value of stocks, commodities, etc.

Actually, resources (e.g., water, land, uranium) do have inherent economic value.

[guenther]

Poor choice of words since "inherent value" is an economic term, but the point still stands. The inherent value exists based on utility, and we could make the exact same utility claim with moral value. (E.g. Saving a life would be inherently morally valuable, while following cultural norms would not be.)

My point is that economic value is a man-made invention, but we still have objective metrics based on human behavior. We could construct a moral equivalent to economics and achieve the same results (though practically I think it would be very challenging to do anything useful).

[CorruptUser]

I guess the best way to measure "good" and "bad" would be measuring the change in value. "Good" would be construction; turning something of lesser value such as rocks, into something of greater value such as a house. "Bad" would be destruction, such as turning the house back into rocks. The measure of how "good" or "bad" might be subjective, but could be measured by how much people are willing to pay for it.

This assumes that the cost of everything are paid for entirely by those purchasing the "product", and everyone benefiting from the "product" pays. For example, if producing the product involved contaminating the local water supply (e.g., mining, farming, power generation), then the company and their insurance company (and in the end, the consumer of the products) have to pay the full cost of the damages done. For "products" whose benefits are spread out, such as police and fire protection, there needs to be a system to force people to pay; if you didn't pay the fireman, the amount of protection you receive is almost the same, but if everyone doesn't pay, there is a crisis.

[Vaniver]

Economic value doesn't inherently exist, but we do have an objective measure for the economic value of stocks, commodities, etc.

What is that?

Because if it's "the price" it should be obvious that the price is subjective. Prices are ways of equalizing subjective values and transmitting information, but the price of something is not objective and unchanging- it is extremely sensitive to location, time, and the rest of the world.

Actually, resources (e.g., water, land, uranium) do have inherent economic value.

What? No, they don't. How would you know what their inherent value is?

Anything has value because a thinking agent places value on it. The value, since it is a judgment call by a thinking agent, does not exist before that thinking agent comes along. It isn't a demonstrable property of the thing.

You can say water is wet, that it has a certain density, that it's incompressible, that it turns into hydrogen and oxygen when you run a current through it- but you can't put a dollar value on it without context.

"Good" would be construction; turning something of lesser value such as rocks, into something of greater value such as a house.

Construction does not make something more valuable. For example, it is unpatriotic to feed bread to pigs.

[Dark567]

Actually, resources (e.g., water, land, uranium) do have inherent economic value.

What? No, they don't. How would you know what their inherent value is?

Anything has value because a thinking agent places value on it. The value, since it is a judgment call by a thinking agent, does not exist before that thinking agent comes along. It isn't a demonstrable property of the thing.

Do you mean absolutely nothing has inherent value or just objects don't have intrinsic value? The former view is not uncontroversial, there is wide disagreement on whether or not intrinsic value exists. Utilitarians argue that pleasure(or well-being, or satisfaction) has inherent value, as an example.

[guenther]

Because if it's "the price" it should be obvious that the price is subjective. Prices are ways of equalizing subjective values and transmitting information, but the price of something is not objective and unchanging- it is extremely sensitive to location, time, and the rest of the world.

What price did Google stocks open at today? I'm pretty sure the answer is independent of the observer. That's an objective metric even if it's a measurement of a large aggregation of subjective takes on value. Economics has lots of these. In the same way we could have objective metrics on morality.

Anything has value because a thinking agent places value on it. The value, since it is a judgment call by a thinking agent, does not exist before that thinking agent comes along. It isn't a demonstrable property of the thing.

You can say water is wet, that it has a certain density, that it's incompressible, that it turns into hydrogen and oxygen when you run a current through it- but you can't put a dollar value on it without context.

I agree with this. The value of gold is really a measurement of us and how we interact with gold, but that doesn't mean we can't objectively measure it. Someone simply has to define a metric that's independent of the observer.

Having said that, I'm really just making an argument over what is possible. I don't think we typically regard morality in this way, and I'm not sure it'd be very useful anyway.

[CorruptUser]

"Good" would be construction; turning something of lesser value such as rocks, into something of greater value such as a house.

Construction does not make something more valuable. For example, it is unpatriotic to feed bread to pigs.

The end value was less, therefore it was destruction, not construction.

[Vaniver]

Do you mean absolutely nothing has inherent value or just objects don't have intrinsic value? The former view is not uncontroversial, there is wide disagreement on whether or not intrinsic value exists. Utilitarians argue that pleasure(or well-being, or satisfaction) has inherent value, as an example.

I mean that value is created by the judgments of intelligent actors. Pleasure, then, has value because intelligent actors value it. Whether or not you want to call that "inherent value" does not trouble me- I will just point out that without an intelligent actor, it is difficult to construct a meaning for 'pleasure.'

What price did Google stocks open at today? I'm pretty sure the answer is independent of the observer.

Note that you used the word "today" not "September 28, 2010." The price of Google stocks changes from moment to moment, instead of being set in stone.

Though a certain price is listed on the NYSE for Google stock at a certain time, it is not clear to me that that price being unambiguous makes it objective. Certainty and uniformity is space and across viewers are necessary conditions, but don't seem sufficient.

The end value was less, therefore it was destruction, not construction.

My point was that redefining construction and destruction in that way is sometimes counterintuitive. If you are comparing start value and end value, what you're interested in is profit, not construction or destruction. Carmakers can, in the process of constructing millions of cars, destroy billions of dollars in value- and so while using symmetric language may have literary benefits I am unconvinced it has clarity benefits. We've already got "profitable" and "unprofitable" floating around, and it seems better that people get over their confusions about those words than borrow other words to take their place.

[guenther]

Note that you used the word "today" not "September 28, 2010." The price of Google stocks changes from moment to moment, instead of being set in stone.

Just like temperature.

Though a certain price is listed on the NYSE for Google stock at a certain time, it is not clear to me that that price being unambiguous makesit objective. Certainty and uniformity is space and across viewers are necessary conditions, but don't seem sufficient.

What is sufficient? I think all that's required is that the metric is independent of the observer. Certainty isn't required since all measurements will have sources of errors. Sometimes the best we can do is estimate an objective quantity.

[Vaniver]

Just like temperature. ... What is sufficient? I think all that's required is that the metric is independent of the observer. Certainty isn't required since all measurements will have sources of errors. Sometimes the best we can do is estimate an objective quantity.

I think you may be right here- and so I will fall back to the consequentialist claim that you generally get better intuition by thinking of prices as subjective in time and place, though you are correct that prices are the most objective measure of subjective value. I have strong qualms about labeling something objective if it depends on something subjective, but those qualms may be misplaced.

[guenther]

I will fall back to the consequentialist claim that you generally get better intuition by thinking of prices as subjective in time and place, though you are correct that prices are the most objective measure of subjective value.

This is an interesting point. I think a lot of people get caught up in seeking "truth" where truth is really poorly defined. So we would do better to seek out useful ideas (e.g. ones that make intuitive sense).

Is value objective? We can define it that way. Is value subjective? We can define it that way. There is no right answer. Or rather, the right answer is however we want to define it. The same is true with beauty. And the same is true with morality. These are all invented concepts, so we get to decide whether the "right" answer is defined subjectively or objectively.

With regards to morality, it seems that it's useful to treat it like it's objective even though the objective metric is poorly defined. My guess is that this has to do with the fact that's it's easier to effect behavior change in others when you have the "right" answer rather than merely a loud opinion. And thus out of usefulness comes a belief in truth.

[Hedonic Treader]

I mean that value is created by the judgments of intelligent actors. Pleasure, then, has value because intelligent actors value it.

It seems to me that neither agency nor semantic intelligence are necessary here. Imagine you have an immobilized dog (ie. without agency and semantic intelligence) - would it make sense to say that there is a difference in value between pleasure and pain of that being? Unless my theory of mind is wrong about the nature of such animals, I think from the dog's perspective, there is a huge and ontologically real value difference, even though it could obviously not use the term. (Materialist eliminativists may disagree, but I never managed to properly wrap my head around their position.)

Do you mean absolutely nothing has inherent value or just objects don't have intrinsic value?

Unless one rejects physicalism, brains in all their possible states are "just objects", too. If you assume that there's a value difference between your own brain being in agony and your own brain being in a hilarious fit of laughter, plus physicalism, you must also assume that physical objects can have intrinsic value. (I think it's hard to disagree that there's a value difference; this is one of the few propositional statements where threatening someone with torture and noting how their answers change would actually tell us something about the truthfulness of these statements.)

[Hedonic Treader]

With regards to morality, it seems that it's useful to treat it like it's objective even though the objective metric is poorly defined. My guess is that this has to do with the fact that's it's easier to effect behavior change in others when you have the "right" answer rather than merely a loud opinion. And thus out of usefulness comes a belief in truth.

The obvious problem here is that a behavior change in others is only useful if it is also true that such a behavior change is desirable, and for something to be desirable in comparison to something else, the existence of value must be true.

[Dark567]

With regards to morality, it seems that it's useful to treat it like it's objective even though the objective metric is poorly defined. My guess is that this has to do with the fact that's it's easier to effect behavior change in others when you have the "right" answer rather than merely a loud opinion. And thus out of usefulness comes a belief in truth.

The obvious problem here is that a behavior change in others is only useful if it is also true that such a behavior change is desirable, and for something to be desirable in comparison to something else, the existence of value must be true.

Right, value must be true for for something to be desirable in comparison to something else. There is the possibility that value isn't true at all though.

Unless one rejects physicalism, brains in all their possible states are "just objects", too. If you assume that there's a value difference between your own brain being in agony and your own brain being in a hilarious fit of laughter, plus physicalism, you must also assume that physical objects can have intrinsic value. (I think it's hard to disagree that there's a value difference; this is one of the few propositional statements where threatening someone with torture and noting how their answers change would actually tell us something about the truthfulness of these statements.)

Huh. I haven't thought of it this way before, slipped my mind(and I am usually so good about recognizing human reactions as physical brain states). The idea that a certain combination of atoms and electrons is more valuable than some other combination of electrons(especially when it is these very atoms and electrons that are deciding value)..... seems impossible. I guess its just one more piece of evidence for nihilism.

[Vaniver]

Or rather, the right answer is however we want to define it.

The answer that agrees with our definitions is whatever we define it to be, yes. But I think there are other ways to measure answers- like how well they predict reality- that are more important.

With regards to morality, it seems that it's useful to treat it like it's objective even though the objective metric is poorly defined. My guess is that this has to do with the fact that's it's easier to effect behavior change in others when you have the "right" answer rather than merely a loud opinion. And thus out of usefulness comes a belief in truth.

I'm not sure I agree here. While you'll get people that overvalue certainty, you have a hard time convincing people your "right answer" is more than a "loud opinion." Consequentialist claims- "my opinion leads to X, Y, and Z" seem convincing if they independently desire X, Y, and Z, or respect them in some way.

But you're right that people (like myself) apply at least part of our subjective moralities to others, and that's necessary to function as a society. Indeed, how much that occurs is part of morality itself.

It seems to me that neither agency nor semantic intelligence are necessary here. Imagine you have an immobilized dog (ie. without agency and semantic intelligence) - would it make sense to say that there is a difference in value between pleasure and pain of that being? Unless my theory of mind is wrong about the nature of such animals, I think from the dog's perspective, there is a huge and ontologically real value difference, even though it could obviously not use the term. (Materialist eliminativists may disagree, but I never managed to properly wrap my head around their position.)

The relevant intelligence when we're discussing pleasure and pain is the ability to feel pleasure or pain. Does it make sense to talk about the pleasure or pain felt by a tree? By a rock? By a collection of water molecules?

It may be valuable to talk about what a tree 'values', but those things will be very tree-centric and on a very limited scale. I cannot think of anything associated with a rock that I am comfortable calling "values" in the same discussion as discussing human values.

That is to say, if you can conceive of something as having a perspective besides the perspective you project onto it, it is relevant to talk of its values. But to say "a rock is intrinsically valuable" seems to me to be an incomplete sentence- who values the rock? When discussing the dog, it is clear who directly values its pleasure and pain, and then indirect value of its pleasure or pain occurs in other intelligent beings.

(I am fine with replacing 'agent' with 'being' so long as 'intelligent being' can be measured- which seems much easier to do with agents.)

The idea that a certain combination of atoms and electrons is more valuable than some other combination of electrons(especially when it is these very atoms and electrons that are deciding value)..... seems impossible.

Hardly. I prefer drinking liquid water in a narrow temperature range, even though you could speed up the same atoms by just a few percent and I would consider it undrinkable and the relevant object being "just" an assemblage of constituent particles and I am "just" an assemblage of constituent particles. Because I am a rather fine assemblage of constituent particles, if I say so myself.

There is more meaning in a file that plays Beethoven on speakers when run and another file which contains the same number of 0s and 1s arranged randomly. If you want to measure that quantitatively, you can start with entropy (but generally meaning requires understanding- back to intelligence!).

[Dark567]

Hardly. I prefer drinking liquid water in a narrow temperature range, even though you could speed up the same atoms by just a few percent and I would consider it undrinkable and the relevant object being "just" an assemblage of constituent particles and I am "just" an assemblage of constituent particles. Because I am a rather fine assemblage of constituent particles, if I say so myself.

There is more meaning in a file that plays Beethoven on speakers when run and another file which contains the same number of 0s and 1s arranged randomly. If you want to measure that quantitatively, you can start with entropy (but generally meaning requires understanding- back to intelligence!).

Sorry, I guess what I meant to say that it doesn't seem like one combination is more intrinsically valuable than another. The reason you prefer your water colder is because when you drink it, the electrons/atoms in your brain arrange themselves differently than if you were to drink it when it was warmer. It is the comparison of these two brain states I am really questioning, why should we prefer the brain state of "I like this water" compared to "this water is too warm" or maybe to rephrase, why is one combination more valuable than the other.

[Vaniver]

It is the comparison of these two brain states I am really questioning, why should we prefer the brain state of "I like this water" compared to "this water is too warm" or maybe to rephrase, why is one combination more valuable than the other.

That question is answered by "Who are we?"

Because if we are ensembles far removed from the atoms in question, it's not clear we should immediately care. But if my ensemble includes the atoms in question, then I have a strong motive for preferring.

[Hedonic Treader]

Right, value must be true for for something to be desirable in comparison to something else. There is the possibility that value isn't true at all though.

Well... I see it like this: Either there is value, or there isn't. And if there is, we can either identify and causally affect it, or we can't.

If there is no value, then there is no value in believing true propositions rather than false propositions. In this case, there is no value in (correctly) believing that there is no value. Neither is there value in adjusting decision-making to the fact that there is no value. Of course, this would violate epistemic rationality, but if there is no value, then there is no value in epistemic rationality.

If there is value but we cannot identify and causally affect it, then we cannot gain value by correctly believing that we cannot identify and causally affect it (because that would imply that we can causally affect identifiable value by believing something true). Again, epistemic rationality would not allow us to gain value.

If there is value and we can identify and causally affect it, then there is value to be gained by identifying and causally affecting it through decision-making. So from the perspective of instrumental rationality, we shoud assume a priori that there is value, and that we can identify it using epistemic rationality as a tool and affect it through decision-making (ethics). An abstract descriptive framework of how physical structures constitute value would then be most helpful (hence this thread).

Or maybe physicalism is false.

[guenther]

The obvious problem here is that a behavior change in others is only useful if it is also true that such a behavior change is desirable, and for something to be desirable in comparison to something else, the existence of value must be true.

The existence of people valuing it must be true.

Well... I see it like this: Either there is value, or there isn't. And if there is, we can either identify and causally affect it, or we can't.

When something is valuable, it just means that people value it. So if we can demonstrate that people value gold then gold has value. But the value is not an intrinsic property of the gold, but rather it's a property of people and how we interact with gold. So there is value, but it's value in the context of people. And we can cast moral value in the exact same way if we want.

The answer that agrees with our definitions is whatever we define it to be, yes. But I think there are other ways to measure answers- like how well they predict reality- that are more important.

How do we predict the reality of morality? I claim we can't because reality of morality is poorly defined, or in other words it's however we define it. When claims can be measured against reality, then they're truth is better defined.

But you're right that people (like myself) apply at least part of our subjective moralities to others, and that's necessary to function as a society. Indeed, how much that occurs is part of morality itself.

I agree, but I'm saying that when people apply their subjective moralities to others, they adopt better rhetoric than "because it feels right to me". And in my experience people will often flat out make claims like "Because it's the right thing to do!" Sometimes people will put more substance behind the claim, i.e. "This will lead to X, Y, and Z", but then X, Y, and Z are in and of themselves poorly defined terms like "equality" or "harm".

So my claim is that we often cast things with poorly defined truth content as if they're true because it makes a more compelling case when trying to motivate behavior in others. And this process actually makes people believe it's true, which of course makes more people present it as truth. In other threads I've described this process as casting wisdom as knowledge. A certain type of behavior is useful to the group, and we store that wisdom in the culture as statements of truth. This is my pet theory on how religion works, but I think this mechanism permeates virtually all aspects of our culture.

[krazykomrade]

There is the possibility that value isn't true at all though.

I have no idea what this means. Truth and falsity are properties of propositions, not concepts. Do you mean that it is possible that objective value doesn't exist? Or that intrinsic value doesn't exist? To deny that value, itself, is a concept without content is almost trivially false.

Along those same lines, I'm having difficulty following the rest of your post, Hedonic Treader. To suppose that there is subjective value, and that we can identify and causally affect it, does not entail that there is value to be gained by doing so, as that would depend on one's subjective values! And if there is objective value (say, pleasure), then we might be able to identify it but we certainly can't causally affect it, as it is objectively valuable! We could causally affect the tokens of that value, that is, change the world in ways that reflect or promote this value, but that doesn't entail that there exists a framework for doing so, as moral situationalism might well be correct form of moral realism.

Most moral people are so without having any framework to track objective value, just as most folks who use good grammar don't have a theory of grammar; and folks who (claim to) believe in and use moral frameworks, monist or pluralist, utilitarian or deontological, have been shown to be no more moral than those whose don't. Perhaps that's because we haven't discovered the "correct" one yet, but that seems awfully implausible.

Also, I don't see how instrumental rationality would dictate that we ought to believe in moral realism (or objective, or intrinsic values), as casually interacting with objects of my own subjective value seems no less desirable or preferable than causally interacting with objects of intrinsic value. Or am I misunderstanding what you meant by value?

EDIT:

So my claim is that we often cast things with poorly defined truth content as if they're true because it makes a more compelling case when trying to motivate behavior in others. And this process actually makes people believe it's true, which of course makes more people present it as truth.

That seems about right, but that also leads to the question of what exactly truth is, or if the concept of truth itself is just one of these notions.

[Dark567]

It is the comparison of these two brain states I am really questioning, why should we prefer the brain state of "I like this water" compared to "this water is too warm" or maybe to rephrase, why is one combination more valuable than the other.

That question is answered by "Who are we?"

Because if we are ensembles far removed from the atoms in question, it's not clear we should immediately care. But if my ensemble includes the atoms in question, then I have a strong motive for preferring.

You have a preference for the brain state of "I prefer X" over "I don't prefer X". Isn't that just stating "I prefer what I prefer"? Or I guess if you make the conscious decision to prefer your brain state "I prefer X" your creating a new brain state of "I prefer the brain state 'I prefer X'". Godelianess ensues.

I guess I am still not convinced why I should value one arrangement state over another, particularly when those states are preferences.

There is the possibility that value isn't true at all though.

I have no idea what this means. Truth and falsity are properties of propositions, not concepts. Do you mean that it is possible that objective value doesn't exist? Or that intrinsic value doesn't exist? To deny that value, itself, is a concept without content is almost trivially false.

Errr... yeah that wasn't nearly as clear as I could have been. "It is possible that intrinsic value doesn't exist", would have been a better choice of words.

[guenther]

That seems about right, but that also leads to the question of what exactly truth is, or if the concept of truth itself is just one of these notions.

I think of well-defined truths as those which have a way to test for correctness. So in math and logic we can have true statements if we grant certain axioms. And the sciences reveal truth to us about the physical world. When there is no test for correctness, we have no way to measure if a claim is true or false. This is what I mean when I say that a claim has a poorly defined truth content.

[Dark567]

That seems about right, but that also leads to the question of what exactly truth is, or if the concept of truth itself is just one of these notions.

I think of well-defined truths as those which have a way to test for correctness. So in math and logic we can have true statements if we grant certain axioms. And the sciences reveal truth to us about the physical world. When there is no test for correctness, we have no way to measure if a claim is true or false. This is what I mean when I say that a claim has a poorly defined truth content.

Well you can test for logical consistency, which isn't so much a test for truth as a test for "this idea isn't so bad that its absolutely impossible". Then again there is no real such thing as a test for correctness, just testing for falseness over and over and if its never false eventually starting to believe it to be true.

[guenther]

Well you can test for logical consistency, which isn't so much a test for truth as a test for "this idea isn't so bad that its absolutely impossible". Then again there is no real such thing as a test for correctness, just testing for falseness over and over and if its never false eventually starting to believe it to be true.

Then maybe there ultimately is no truth about the real world. But treating things as if they're true is still useful. Like for instance if a theory has been tested over and over and it has yet to be falsified. I'm happy to regard the scientific method as a way to test for truth (with the understanding that any one experiment doesn't directly prove truth).

[Hedonic Treader]

And if there is objective value (say, pleasure), then we might be able to identify it but we certainly can't causally affect it, as it is objectively valuable! We could causally affect the tokens of that value, that is, change the world in ways that reflect or promote this value, but that doesn't entail that there exists a framework for doing so, as moral situationalism might well be correct form of moral realism.

Yes, I meant causally affecting the tokens of that value.

Most moral people are so without having any framework to track objective value, just as most folks who use good grammar don't have a theory of grammar;

Many people (myself included) have an intuitive tendency to empathize with fictional characters while failing to empathize with out-of-sight suffering of other real-world sentients, or sentients who have an out-group label attached to them, or sentients who are ugly etc. There is a case both for explicit epistemic rationality and instrumental rationality in ethics. A descriptive framework of what good and bad actually mean in a physical universe seems a meaningful goal in this case, if it is possible. There's no expectation on my part that the accuracy of such a framework would be perfect. Nevertheless, formal epistemic methods can hardly ever be proven to be perfect, but that doesn't mean they're inferior to folk wisdom (think folk wisdom vs. scientific method).

and folks who (claim to) believe in and use moral frameworks, monist or pluralist, utilitarian or deontological, have been shown to be no more moral than those whose don't.

You do realize that statement pre-supposes that some people can be seen objectively as more moral than others? According to what framework then? Their own explicitly defined ones?

Also, I don't see how instrumental rationality would dictate that we ought to believe in moral realism (or objective, or intrinsic values), as casually interacting with objects of my own subjective value seems no less desirable or preferable than causally interacting with objects of intrinsic value. Or am I misunderstanding what you meant by value?

If physicalism is true, you are a physical object, and your properties are objective facts about the physical world. Under this premise, where does your subjective value come from? I see no other way than to assume that your brain (in certain states) is a physical object of intrinsic value then. By "causally interacting with objects of my own subjective value", you are ultimately "causally interacting with objects of intrinsic value" (your brain in various states).

[krazykomrade]

Then maybe there ultimately is no truth about the real world. But treating things as if they're true is still useful.

But isn't that what you earlier pointed out was what people erroneously do with things of questionable truth value? I mean, to call things true which, strictly speaking, you can't know to be, simply because they are useful; doesn't that make the notion of an objective truth, bearing on all, just such an illusion which is used to make a more compelling case (to yourself, if not others!)?

Many people (myself included) have an intuitive tendency to empathize with fictional characters while failing to empathize with out-of-sight suffering of other real-world sentients, or sentients who have an out-group label attached to them, or sentients who are ugly etc. There is a case both for explicit epistemic rationality and instrumental rationality in ethics. A descriptive framework of what good and bad actually mean in a physical universe seems a meaningful goal in this case, if it is possible. There's no expectation on my part that the accuracy of such a framework would be perfect. Nevertheless, formal epistemic methods can hardly ever be proven to be perfect, but that doesn't mean they're inferior to folk wisdom (think folk wisdom vs. scientific method).

I'm not sure why you keep bringing in the notion of a physical universe, and I'm actually not to sure what you're saying here, at all. I find the dichotomy between folk wisdom and the scientific method radically disanalgous to morality and ethical frameworks. For one, the scientific method clearly works better than folk wisdom (in many/most applications), but ethical frameworks don't seem to morally improve those who know or use them. In fact, the only means by which to evaluate ethical frameworks (other than internal consistency) is by comparison to our own moral intuitions (folk morality)! How else would you be able to know if a ethical framework was giving you "correct" answers?

In short, I'm still not sure why you think moral realism is to be preferred on instrumentally rational grounds, and supposing it is, I'm still not sure why you think moral situationalism is not the case (or are you merely rejecting it on instrumentally rational grounds as well? That'd be kinda odd, as the failure of monism and pluralism to be instrumentally rational is a common arguement in *favor* of moral situationalism!)

If physicalism is true, you are a physical object, and your properties are objective facts about the physical world. Under this premise, where does your subjective value come from? I see no other way than to assume that your brain (in certain states) is a physical object of intrinsic value then. By "causally interacting with objects of my own subjective value", you are ultimately "causally interacting with objects of intrinsic value" (your brain in various states).

Why should I suppose physicalism is true? Kant did a fine job of showing why there's no reason to suppose it is. If it were, I'm not sure the notion of objective value even make sense. I mean, what would it even mean for a collection of of matter to have objective value, independent of other matter thinking it does? To say that my brain is of value because it is the basis of all my other values, is to say that it is subjectively valuable, that is, it is only valuable in terms of me and my values!