Philosophical Vegan Forum

About a week or two ago, Vegan Gains hosted a debate over the validity of Name the Trait. The participants were Vegan Foot Soldier, Perspective Philosophy, Shadow Starshine, and somebody named Bryn. It seemed that one of the big issues between the debaters was a lack of a rigorous definition for the term "trait equalizable" in premise 1 of the revised Name the Trait argument.

I have a math background and tried to write up a rigorous definition of "trait equalizable" using set theory. The definition I came up with is inspired by Singer's Marginal Cases Argument. After creating the definition, I also performed some analysis on what it would mean when applied to Name the Trait, and came up with a few criticisms. I was wondering if anybody on this forum would be interested in looking at the analysis?

I created a story on medium.com because the definition and analysis seemed a little lengthy for a forum post. If anybody is interested in taking a look at it/providing feedback, it can be found at: [link]

Any feedback would be greatly appreciated! As I said, I have a math background, and enjoy applying math to real world problems. I also became interested in Singer's argument, so it seemed like an interesting argument to translate into a mathematical context. But, it isn't always easy to get feedback on this sort of thing, so I appreciate anybody that is willing to take some time to look at what I've written.

Hi Priv, we don't normally allow links for new members. However, I assure you it isn't too long for a forum post. We have some very long posts here. Please feel free to copy it here as a reply.

That said, there are different ways to interpret "trait equalizable", but the only thing that's needed is possibility in a parallel hypothetical copy of the universe rather than the actual universe.
I critiqued the metaphysics of a stronger interpretation in the past regarding "exploding cows" and temporal issues. You don't even have to make the kinds of arguments PP and SS make, there are literal impossibilities in physics itself with a strong interpretation.

A very permissive interpretation like that one that only looks at hypothetical copies of the situation makes #NTT subject to attacks critiquing hypothetical or conceptual moral systems as being add odds with morality in reality, but these criticisms must presume that morality is in fact NOT a concept but a thing of metaphysical substance (like "sin" as a material condition).

#NTT could possibly be stronger and less prone to this red herring by incorporating such a premise and thereby proving the trait equalizability premise, but it seems like only a small fringe are really pushing this argument so I'm not sure if it's worth it or not because it might make the argument longer.

Hello brimstone,

Thank you for taking a look! I understand that "trait equalizable" may have several definitions, and was going to look at the parallel universe interpretation as well. But, I was also interested in whether or not there might be problems with my interpretation. I think that some of my criticisms, primarily the ideas about compound properties, might be applicable to a parallel universe view as well, primarily because traits seem to be analyzed in an iterative fashion in the parallel universe approach as well.

But, I wanted to get an idea on whether or not there may be flaws in the analysis of the definition I proposed because I think it's a little easier to understand. In any event, I've copied the paper below:

--- Preliminaries ---

The term “trait-equalizable” appears in the formal Name the Trait argument, however the term does not appear to be rigorously defined. This paper is proposing a rigorous definition for trait equalization that is consistent with the claims made in Peter Singer’s Marginal Cases Argument. I am going to use the IEP version of the argument (https://www.iep.utm.edu/anim-eth/#SH3a):

S1. In order to conclude that all and only human beings deserve a full and equal moral status (and therefore that no animals deserve a full and equal moral status), there must be some property P that all and only human beings have that can ground such a claim.
S2. Any P that only human beings have is a property that (some) human beings lack (e.g., the marginal cases).
S3. Any P that all human beings have is a property that (most) animals have as well.
S4. Therefore, there is no way to defend the claim that all and only human beings deserve a full and equal moral status.

Before proceeding, I also want to clarify one thing. The above formulation of Singer’s argument refers to moral status. According to SEP (https://plato.stanford.edu/entries/grou ... al-status/), moral status can be grounded only by an entity’s intrinsic value. So, I will attempt to create a set theoretic approach to describe intrinsic value, and use this as the basis for trait equalization.

--- Introduction ---

In “Animal Liberation”, Peter Singer explains that certain intrinsic properties, such as skin color, have wrongly been used as the basis for discrimination. He also acknowledges that there are certain properties, such as intelligence and moral agency, that are present in some humans, but not in nonhuman animals, that may account for the assertion that a specific human has more intrinsic value than all non-human animals. But, he further argues that not all humans possess such a trait. To make this more concrete, suppose that the intrinsic value of an entity can be described by a set of properties. For example:

Intrinsic value of adult: { sentience, moral agency, … }
Intrinsic value of human infant: { sentience }
Intrinsic value of pig: { sentience }

The assertion that there are some humans (infants) that do not possess a property P that distinguish them from animals (the pig) can be formally verified by examining a set of morally relevant properties that contribute to the intrinsic value of an entity. With this in mind, I’m proposing a set based approach that can be used to evaluate the intrinsic value of an entity. When I say morally relevant, I’m acknowledging Singer’s assertion that certain properties do not contribute to the intrinsic value of an entity, therefore they should not be included in the set that corresponds to an entity’s intrinsic value.

For this formalization to be useful, we have to be willing to accept the presumption that a set of properties that can contribute to an entities intrinsic value can be constructed. For example, maybe the set is { sentience, moral agency, empathy }. How this set is constructed is not addressed here — it’s a question for the philosophers. But, I think that Singer would agree that such a set exists. Second, we must also presume that an entity can be described by a set of properties, some of which contribute to its intrinsic value, some of which do not.

One important note: when I say that a property contributes to the intrinsic value of an entity, it can do so in a positive or negative fashion. For example, since the properties that are used to describe the intrinsic value of an adult human is different from the set that is used to describe the human infant, this paper will draw no conclusions about which is of more intrinsic value. It’s possible that moral agency contributes negatively to the intrinsic value of the adult and to use this to conclude that the infant has more intrinsic value. However, since the set of properties used to describe both the pig and human infant are identical, we will presume that the intrinsic value of each is also identical.

With that in mind, I’m going to propose the following definitions:

1) Let M represent the set of all morally relevant properties that can contribute to an entity’s intrinsic value. For example, M = { sentience, sapience, empathy, moral agency, … }.
2) Each entity can be represented by a set of properties. For example, a pig can be represented by a set that includes sentience, various physical properties, etc. An adult human can be represented by the set that includes moral agency, intelligence, sentience, empathy, various physical traits, etc. We will call the set of properties that can be used to describe an entity as the entity’s traits.
3) We will define the intrinsic value of an entity as the intersection of the entity’s traits and M.
4) We will say that two entities are trait equalizable if the intrinsic value of each entity is identical.
5) We will call any process that can be used to properly determine if two entities are trait equalizable a trait equalization process.

--- Example of TEP ---

Suppose we define M = { sentience, moral agency, empathy }. We wish to evaluate whether or not a pig is trait equalizable to a specific human. We’ll try two humans: an adult and an infant. First, we define the traits of each entity:

Pig traits = { sentience, tough skin, … }
Infant traits = { sentience, soft skin, … }
Adult traits = { sentience, moral agency, empathy, not so soft skin but not tough either, … }

Next, we calculate the intrinsic value of each: the intersection of an entity’s traits with M:

Pig intrinsic value = { sentience }
Infant intrinsic value = { sentience }
Adult intrinsic value = { sentience, moral agency, empathy }

Then we compare the intrinsic values of each entity. Note that the pig and infant are trait equalizable, but the pig and the adult are not. Suppose there is a correspondence between an entity’s traits and whether or not the entity has moral status. Since some infants are trait equalizable to pigs, if we claim that infants have moral status, we must also conclude that pigs should also have moral status.

One thing to note: the calculation of an entity’s intrinsic value is an atomic operation in this example. In other words, we do not calculate intrinsic value in an incremental fashion. I’ll have a bit more to say about this in the section titled “Metamorphosis”.

--- Deriving Contradictions ---

Trait equalization may be useful when trying to determine our obligations to an entity. If our obligations to an entity are grounded by their claim to moral status, and the moral status can be determined by the entity’s intrinsic value, the entity’s intrinsic value corresponds to our obligations.

For example, if an infant has full moral status, and the grounds for this full moral status is intrinsic value { sentience }, then any other entity with the intrinsic value of { sentience } should also have full moral status, and a claim to the same obligations as the infant. With that said:

Premise 1: We have an obligation to not harm any entity that has full moral status.

If we accept this premise, we must also accept that we have a moral obligation not to harm an entity with intrinsic value of { sentience }. So, suppose we agree with premise 1. We can use the trait equalization process to demonstrate that both the infant and the pig have equal intrinsic value. If they do, we have a moral obligation not to harm either. Any other conclusion would be inconsistent in this framework.

But, it is very important to note that Premise 1 assumes the obligation not to harm is based on moral status, and thus, intrinsic value. If the grounds for this obligation includes reasons other than the intrinsic value of an entity, Premise 1 is in dispute, and we can make no claims to inconsistency.

--- Metamorphosis ---

When we calculate the intrinsic value of an entity, the evaluation is performed atomically. However, it is possible to view it as a process of elimination, where we remove properties one by one from our set of traits. For example, suppose M = { sentience, moral agency } and a human’s traits are: { sentience, moral agency, hair color, eye color }. We can perform a series of incremental changes:

{ sentience, moral agency, hair color, eye color } — is eye color in M? No, so remove it.
{ sentience, moral agency, hair color } — is hair color in M? No, so remove it.
{ sentience, moral agency } — at this point, the remaining traits are in M, so we have calculated the intrinsic value of the adult.

There is nothing problematic about this approach if we keep in mind the we do not attempt to evaluate trait equalization until we have completed the iterative process. To show why this is important, I’m going to re-visit Singer’s argument.

--- Marginal Cases Revisited ---

Singer states that we cannot claim that every human has full moral status and also deny animals a full moral status unless there is a single property P present in every human to ground such a claim. This is actually inaccurate. There may be several such properties. For example, suppose there exist properties P1 and P2 that contribute to intrinsic value, and:

intrinsic value of infants: { sentience, P1 }
intrinsic value of all other humans: { sentience, P2 }
intrinsic value of pig: { sentience }

Further suppose that P1 and P2 are not present in any animal. Then there are two properties that can ground the claim that every human has moral status while denying this claim to animals.

Why is this important? Recall, we previously described the intrinsic value of human adults and infants as:

Infant intrinsic value = { sentience }
Adult intrinsic value = { sentience, moral agency, empathy }

If the intrinsic value of an entity corresponds to an obligation not to harm the entity, it is tempting to conclude that the property sentience corresponds to this obligation because it is the only property that an infant and an adult have in common. It would also be tempting to claim that we could arrive at this conclusion by process of elimination.

However, this is not justified by the framework laid out in this paper. Suppose that P1 and P2 exist. If we want to examine P1 to determine if it is the reason for obligation not to harm, we could conclude no, because it is not present in “all other humans”. If we want to examine P2 to determine if it is the reason for the obligation not to harm, we could also conclude no, because it is not present in infants. In this way we might claim that sentience is the only reason every human has moral status. However, in the first example, it was shown that every human could have moral status because either P1 or P2 is present in every human.

--- Other Compound Properties ---

Suppose we include the property “potential for moral agency” in the list of morally relevant properties. One of the arguments I’ve run across is that if this is the property that grounds moral status, then a non-sentient fetus must also have moral status. But, let’s look at some intrinsic values:

non-sentient fetus: { potential for moral agency }
infant: { sentience, potential for moral agency }
animal: { sentience }

It should be clear that all of these entities have different intrinsic values. The non-sentient fetus lacks sentience, whereas the animal lacks the potential for moral agency. Only the infant has both properties. In this case, the presence of two properties may be the grounds for moral status.

--- Name the Trait ---

I’ve run into several types of arguments that seem to be inspired by name the trait that I have problems with. I’m going to describe how my definition of trait equalization is incompatible with these types of arguments.

First, there is the “human slowly transforms into a pig” argument. The ideas laid out in this paper do not support such an argument. It’s important to keep in mind that this paper’s approach is intended to evaluate the intrinsic value of two entities and compare the intrinsic values for equality. It is not the case that one entity is slowly transformed into another by removal/replacement of specific properties. In fact, only a subset of properties participates in the evaluation of equality, so we’re not even comparing the entities themselves.

A second type of argument is the process of elimination type argument. For example, name a trait that distinguishes an animal from a human with respect to treatment. I say moral agency. Moral agency is not present in infants, so it cannot play a role in making the distinction. I’ve provided a counter example to demonstrate why this may not be the case.

Another problem is the role of non-intrinsic traits. For example, suppose our obligation not to harm a human is grounded in social contract theory. The usual “reductio” is to say, well, the trait is “people capable of participating in a social contract”. While this trait may contribute to the intrinsic value of an entity, it is not the same thing as social contract theory. One is an intrinsic trait, while the other is a branch of normative ethics. There is no logical basis for equating the two.

Finally, the biggest problem with name the trait. Suppose I name the trait of moral agency. In order to examine this trait, we would need to select a human that possesses this trait. For example a human adult. When you compare a human adult with an animal, the human adult is not trait equalizable to an animal because the adult possesses the trait of moral agency. At this point, I suppose we could then point out that infants do not possess the trait of moral agency. This is simply stating what Singer has already stated in the marginal cases argument. So, is there really a need for name the trait? Does it really add any value to the conversation?

--- Potential Problems with the Definition ---

Suppose a bear wanders into a village. I think most people would consider it morally acceptable to kill the bear. But, suppose an infant wandered into a village. We would not consider it morally acceptable to kill the infant. One obvious explanation is when we kill the bear, we are protecting the villagers. This is a non-intrinsic reason. But, let’s take a look at how we might define the intrinsic value of the bear and infant:

intrinsic value of infant: { sentience, lacks moral agency }
intrinsic value of bear: { sentience, lacks moral agency }

Once again both are trait equalizable. But, is it possible the infant has more intrinsic value than the bear? Yes. Suppose the lack of moral agency in the bear is more meaningful because the bear is physically capable of harming the villagers, and less meaningful in the infant because the infant is not capable of harming the villagers. This would mean that, just because two entities are trait equalizable, they still may not have the same intrinsic value. If this is the case, my approach cannot be used to formally evaluate Singer’s argument. In fact, there may a problem with Singer’s argument because the two entities would have different intrinsic value even though the infant does not possess a property P that the bear does not possess.

It is possible to resolve this by adding another property: “physically capable of harming others”. But, it is counterintuitive to include a physical property in the set of properties that contributes to intrinsic value. We would never accept skin color or eye color as a trait that contributes to intrinsic value. Since this evaluation process is dependent on properly defining the set of morally relevant properties, and we may leave out properties that are seemingly unrelated to intrinsic value, how effective is it in determining our moral obligations?

privkeav wrote: ↑Mon Oct 14, 2019 9:11 am I understand that "trait equalizable" may have several definitions, and was going to look at the parallel universe interpretation as well.

It's not really a parallel universe: it's a hypothetical one. While it is a conceptual mirror, it's *nested* in this one.

privkeav wrote: ↑Mon Oct 14, 2019 9:11 amBut, I was also interested in whether or not there might be problems with my interpretation.

For sure there are issues with other interpretations, the same can be said with any argument. Which is why the author recognizes the hypothetical as a correct one, and the wiki covers this too. If there is a way to make this more clear that would be a good update.

Obviously if you take a human Bob and you take a Cow Bessie, and you rearrange the molecules in the human and add more to make him structurally identical to Bessie, then you name the trait "position" as relevant, and you change their locations to occupy the same space you have an explosion and not a cow (nuclear fusion). To avoid that you'd have to remove half the molecules thus either Bob or Bessie has been removed.

To point is that there are any number of physical impossibilities that preclude actually changing two entities traits in totality to being identical. And you also run into the indiscernibility of identicals problem (e.g. that two things that are identical may be the same thing) -- although that's also kind of the point.
In a hypothetical universe this isn't a concern:

Your view affirms a given human is trait-equalizable to a given nonhuman animal while retaining moral value.

They can literally be made the same thing within the hypothetical. And provided you understand trait as to be a property confined within a universe, there aren't any identity issues.
You really should read the Wiki entry. You can skip to here:
wiki/index.php/NameTheTrait_2.0%2B_(off ... impossible

Philosophical Vegan Wiki wrote:Another means of avoiding this issue is a very specific definition of trait that limits traits to those within the reality being considered.

So, that is for the hypothetical reality, the fact that we outside of that reality are hypothesizing about it is not a trait of entities within that reality.

I recommend reading the all of those criticism sections, you'll understand better the context NTT works in. It has been refined over time to remove weaknesses.

I don't tend to use the NTT argument, and I was a big critic of it before it was valid, but it's valid now.
Whether I find it useful in my argumentation or not some people are persuaded by it to acknowledge that moral value is linked to trait differences and in a Socratic sense to break that down and better understand their value systems and the importance of consistency.

privkeav wrote: ↑Mon Oct 14, 2019 9:11 amI think that some of my criticisms, primarily the ideas about compound properties, might be applicable to a parallel universe view as well, primarily because traits seem to be analyzed in an iterative fashion in the parallel universe approach as well.

I'm not sure how that's relevant. I looked over those sections of your piece.

privkeav wrote: ↑Mon Oct 14, 2019 9:11 am--- Other Compound Properties ---

Suppose we include the property “potential for moral agency” in the list of morally relevant properties.

First, that's absurd.
Potential based arguments are not coherent, because potential is meaningless without context. Likewise, "moral agency" as a concept is fraught with issues and needs a lot of clarification.

A single cell shed from your mouth in your spit has the "potential for moral agency", it's just a question of the environment it finds itself in. A normal cell can be converted into a stem cell, into an embryo, and grow into an adult human with so called "moral agency". All of the information is there in the right environment.
So the question that must be answered to make "potential" meaningful is what environment will it be in?
Well, we can answer that negatively too.
Were I to find you sleeping (without agency) and kill you before you woke, based on that environmental certainty of you being murdered, you'd have no more potential for moral agency because dead people don't have it. The fact that your environment guarantees your death and loss of moral agency strips you of potential and thus makes it OK for me to kill you under that system.

Potential becomes mired with self-fulfilling prophecy. As long as people aren't *currently* expressing moral agency, they're fair game to kill because if you succeed at killing them then they had no more potential because of the murderous environment. And hey, if you kill them fast enough (between a thought, without them being able to reflect on it) you'd get a pass too.

Any time somebody brings up "potential" that's a red flag that he or she just hasn't thought carefully about any of these issues.


That said, compound properties aren't an issue for the validity of the NTT argument. A person can easily disagree with P2 by naming a trait or set of traits like that. A person can disagree with P2 for any reason, like the trait "doesn't wear clothes", and agree that if a human lost these things then the human would lose moral value too.
The point of NTT isn't that nobody can disagree with P2, it's to encourage naming the *reasons* for disagreement so they can be addressed.

You seem to have completely missed the plot here.
Other arguments address whether those are good reasons or not once they have been exposed. While you could only value those who wear clothes, there's some argument to be made that this isn't a good value system even if you are consistent about it -- but again, that argument is separate from the NTT argument. Once (and if) a person agrees that wearing clothing isn't a good criterion for moral value, THEN you could come back to NTT and repeat the question asking what is. That's why NTT seem cyclical; because arguments tend to come back to it after a trait has been debunked as a plausible source of moral value. However, if people stick to their guns and are consistent about "wearing clothes" being the basis for moral value then that's the end of it: NTT no longer applies once a person has demonstrated consistency.

privkeav wrote: ↑Mon Oct 14, 2019 9:11 amOnce again both are trait equalizable. But, is it possible the infant has more intrinsic value than the bear? Yes. Suppose the lack of moral agency in the bear is more meaningful because the bear is physically capable of harming the villagers, and less meaningful in the infant because the infant is not capable of harming the villagers. This would mean that, just because two entities are trait equalizable, they still may not have the same intrinsic value.

If ability to harm compounded with lack of moral agency is morally relevant, then you have NOT equalized all morally relevant traits.
Try again with an infant driving around in an assault vehicle and randomly launching missiles.

If a trait like that (poses a danger, etc.) is morally relevant and has not been equalized, then you've completely evading the whole point of the argument.

It's like one of those people who doesn't understand the point of a hypothetical and answers the Trolley problem with "I switch the track half way so the trolley derails and nobody dies".

You seem to somewhat realize that:

privkeav wrote: ↑Mon Oct 14, 2019 9:11 amIt is possible to resolve this by adding another property: “physically capable of harming others”. But, it is counterintuitive to include a physical property in the set of properties that contributes to intrinsic value.

Despite you already saying you thought the infant might have more intrinsic value because of that, what's your fixation on intrinsic value here?
We make decisions based on extrinsic value too. The argument doesn't specify, it's just dealing with moral value of whatever type (intrinsic or extrinsic).

Most people are naturally consequentialists and make decisions at least largely based on outcome.

You could even say they have the same moral value if you want to limit it to intrinsic value, but that's not the end of all decision making and we can find it right to kill something that has the same (intrinsic) moral value and not kill another thing because of outside factors -- like saving lives which add up.

You're really missing the point on a lot of this, so I'm wondering if you've engaged with anybody on this yet or not. It seems you haven't read the wiki.
Where are you getting your information on NTT? Is it jut from the Vegan Gains debate?

Again, I'm not any big advocate of the NTT argument, I use very different arguments, but you don't seem to understand what you're criticizing here.
Let me know if I'm missing anything in your article, I read the sections you said were relevant to the hypothetical use.

First off, I'd like to assure you that I have read the Wiki article, and I understand that the parallel universes are hypothetical in nature. I'll admit that I haven't completely analyzed this approach to the hypothetical, but will attempt to do so at a later date.

That being said, I think you're getting hung up in the details, and missing some general points that I was trying to make. First, my primary opposition to two entities being made "identical" is based on the concept of particulars and universals. If you start with two entities, for example, a human and a cow, they can be viewed as two particulars, and two particulars will always be discernible in some way because they are two different objects. But the set of traits that is used to describe the two particulars can be made identical because the set of traits can be viewed as a universal. If moral value corresponds to the set of traits, the law of identity should be applicable because a universal is basically a singular entity. But, I don't think the law of identity can be applied to two particulars because they will always be discernible in some way.

EDIT: To clarify the above statements, suppose you have two balls that are exactly the same color: red. The two balls will always be discernible because there are two objects, but their color may be identical. We could then apply the law of identity to the color because the color is an attribute that the balls have in common and the color is identical. In the same way, the set of traits used to describe the participants of the NTT analysis can be considered an attribute of each object that is being hypothetically changed.

This may seem to be a nit pick, but, in the case of NTT, suppose the set of traits that is used to describe each object is identical. Next, suppose that the person is asked, "would you kill the human at this point in the process?". If the person is unwilling to kill the "human", is the person unwilling to kill the object because they have objectively evaluated the set of traits, or are they unwilling to kill the object because they know that it was once a human? I believe this is a legitimate question.

Another point of concern is as follows: what is the point of NTT? Is it a process through which traits that are potentially relevant to moral value are analyzed and eliminated one by one? If so, then, the points that I made in the paper about compound traits are valid. Suppose you have a trait of the form T = T1 and T2 (both need to be present for moral value). Further suppose that you use NTT to analyze only T1 and conclude that T1 is not sufficient for moral value. Has T1 been completely eliminated from analysis? Or, is it known that even if T1 has been eliminated from analysis, that it can be part of a compound trait? If it has been completely eliminated from analysis, then a trait of the form T = T1 and T2 may never be posited.

I understand your criticisms of the property "potential for moral agency", but you seem to have missed the point. For example, suppose we have three entities with the following morally relevant traits:

E1: { P1 }
E2: { P1, P2 }
E3: { P2 }

If these traits can be used to describe the moral value of each entity, it should be clear that the sets used to describe the moral value of the three entities are not identical. Since there is no formula for calculating the moral value based on the presence of the traits, nothing else can be said. But, if you have the following two entities:

E1: { P1 }
E2: { P1 }

it is reasonable to conclude that both have the same moral value because each entity has the same morally relevant traits. But, once again, there is a correspondence between the sets used to describe morally relevant traits with moral value. As I've already stated, trying to form a correspondence between the objects while referring to the objects using their original labels is a problem in my opinion. I also think that my definition addresses this problem.

One last thing, I don't have a problem with the validity of the NTT argument itself. However, I do think that the notion of treat equalization can cause problems if not done properly.

I would like clarify one thing. In my response I asked:

what is the point of NTT? Is it a process through which traits that are potentially relevant to moral value are analyzed and eliminated one by one?

I do understand that the actual point of NTT is to test the consistency of one's moral reasoning. I made this comment because there are some that use the NTT argument that may break up a combined trait into components, and then analyze each component. This obviously occurs outside of the actual NTT argument, because once you start to analyze the validity of a trait or components, you're no longer discussing logical consistency.

However, the wiki for NTT includes a list of traits, that I presume, have been "defeated" by the NTT argument. It is not clear whether these traits can be combined with each other, or combined with other traits that are not included in the list to posit a new trait.

Hopefully this makes sense.

You're missing the point. This is a matter I had a contention with before too until Isaac accepted the "within universe" limit of traits.

privkeav wrote: ↑Wed Oct 16, 2019 7:01 pmI don't think the law of identity can be applied to two particulars because they will always be discernible in some way.

They are discernible by US by the conceptual lineage and the label we have given them in *our* universe, but not from within the hypothetical; not if everything else has been changed in that hypothetical.

They ARE the same thing in that hypothetical, they are only different to US outside that hypothetical universe because we've labeled them differently and remember them being derived from something else and having a different hystory in the different hypothetical universes we've imagined.

The "within universe" limit to traits solves that issue entirely. It does, however, present another issue: If you believe moral value is derived from something outside the universe, then in such a claim you'd disagree with P1 because "moral value" would not be a trait as so defined.

privkeav wrote: ↑Wed Oct 16, 2019 7:01 pmIf the person is unwilling to kill the "human", is the person unwilling to kill the object because they have objectively evaluated the set of traits, or are they unwilling to kill the object because they know that it was once a human? I believe this is a legitimate question.

A trait can be "used to be human". Just amend the hypothetical so the creature was never human to begin with; only conceptually liked by occupying the same place in the thought experiment.

privkeav wrote: ↑Wed Oct 16, 2019 7:01 pmAnother point of concern is as follows: what is the point of NTT? Is it a process through which traits that are potentially relevant to moral value are analyzed and eliminated one by one?

Fundamentally, NTT is meant to aid in the refutation of the implicit claim that it's not necessary to justify moral devaluation.
If you succeed in justifying it to yourself or others, then you have gotten to the other side of the argument.

In practice, most opponents offer internally contradictory justifications. NTT is meant to expose this.

If your justification is "moral value comes from wearing clothes and non-humans don't wear clothes" then NTT would question your consistency.
If it turned out you were fine with farming a culture of humans who are nudist, and would oppose farming some animals who wear clothes (perhaps making the case for hermit crabs, etc.) then you would be consistent. At that point, as far as its main advocates go, NTT would no longer be brought up. You won.

It's not that hard to find a trait and bite the bullet on the implications of consistent application.

privkeav wrote: ↑Wed Oct 16, 2019 7:01 pmIf so, then, the points that I made in the paper about compound traits are valid.

No, it's irrelevant. You might as well have made the point about NTT supposedly not being able to address the "wears clothes" trait for moral value: if you're consistent you're consistent, period. NTT isn't an answer to every trait, it's a consistency test.

That's why I don't really use it, because it's easy to get around by being a consistent idiot. All you have to do is stick to your guns regardless of the implications and you "beat" NTT.
Isaac would argue that those victories over NTT are phyrric. He considers those a win because they demonstrate now insane your value system has to be to justify confining and torturing feeling animals for taste pleasure. However, this is mostly only useful when you have an audience for a debate.

privkeav wrote: ↑Wed Oct 16, 2019 7:01 pmHas T1 been completely eliminated from analysis? Or, is it known that even if T1 has been eliminated from analysis, that it can be part of a compound trait?

The latter of course. I don't know where you could have heard otherwise, that's a very silly straw man. Compound traits have never been a problem or controversial as answers to the argument. "Traits" are not discrete quantum of data anyway. Any "trait" is likely to be a combination of factors. Like even the trait "human" has a combination of genetic variants that distinguish us from Chimps; it's not like you have to narrow it down to one gene.

Sorry, didn't see your reply until posting:

privkeav wrote: ↑Thu Oct 17, 2019 7:10 pm I made this comment because there are some that use the NTT argument that may break up a combined trait into components, and then analyze each component. This obviously occurs outside of the actual NTT argument, because once you start to analyze the validity of a trait or components, you're no longer discussing logical consistency.

Yes, because this is just an easier way to do it.
And yes, that's aside from the argument as you noted.

privkeav wrote: ↑Thu Oct 17, 2019 7:10 pmHowever, the wiki for NTT includes a list of traits, that I presume, have been "defeated" by the NTT argument. It is not clear whether these traits can be combined with each other, or combined with other traits that are not included in the list to posit a new trait.

Most people's intuitions compel them to find a "single" or more coherent and eloquent trait, usually something that resonates as meaningful, rather than an ad-hoc hodgepodge of traits to get the result they want.

You can balk at heliocentrism and make a model of the solar system that's geocentric and works too-- that is, if you make the laws of gravity complicated enough with a bunch of special rules. That is not, however, how rational people prefer to view the world.

I would wager that most rational people intuitively understand it to be intellectually dishonest to create an overly complex ad hoc system just to avoid having to admit you were wrong.

If in any sense you treat your moral value system with a fraction of the critical thinking most skeptics prefer to apply to other empirical matters, then I think you have to challenge yourself to reject that kind of thinking and in the very least work on answering the hard questions of WHY all of those arbitrary intersecting traits you needed to get the answer you wanted should be relevant.

So sure, you CAN combine a bunch of traits, make conditions and exceptions, make some traits which only mean anything if other traits are absent, you can make your list a mile long and so complex you need a spreadsheet with thousands of nested logical operators to actually parse it... but is that the kind of person you are? Is that the kind of answer and thought process that satisfies you?

OK, it seems like we're drifting off what I had originally hoped to discuss, which was the way that I would define trait equalizable. I apologize, because I helped to drift off topic.

I know that I wrote about trait equalization in length, but it all can be described as follows:

1. An entity can be described using a set of traits. For example, sentience, physical traits, etc.
2. Some of these traits are relevant with respect to moral value, and some are not.
3. Ignore the traits that are not relevant.

What happens if we were to do this for a couple of examples? Maybe we would get something like the list below:

human adult: { sentience, moral agent }
infant: { sentience }
cow: { sentience }

What can we infer from this? That the traits used to describe the moral value of the cow and the infant are identical. This observation could be used to argue that the infant and the cow have the same moral value.

There is no transformation. There are no exploding cows (which is an observation that I really enjoyed reading, BTW). The approach does not attempt to validate traits as morally relevant, which I think we both agree, is not the purpose of NTT. It is a simple, well defined mathematical operation that can be used to evaluate the set of traits that would describe an object's moral value.

I know that I do not understand your definition in its entirety. And, I apologize if it seemed like I was suggesting that your argument cannot be defended - I was not. But, what I'm interested in is if my definition can escape criticism.

Can you provide a criticism for my definition? The criticisms you originally made seem to be about the examples I chose, and my knowledge of NTT, not about the general approach I was trying to describe.

As to point 3, you could specify that in the argument.

Like:

P1) If your view affirms a given human is trait-equalizable with respect to morally relevant traits to a given nonhuman animal while retaining moral value, then your view can only deny the given nonhuman animal has moral value on pain of P∧~P.

P2) Your view affirms a given human is trait-equalizable with respect to morally relevant traits to a given nonhuman animal while retaining moral value.

C) Therefore, your view can only deny the given nonhuman animal has moral value on pain of P∧~P

The argument isn't currently framed like that, but I do not see that as being a problem.

Keep in mind this also would have no problem with compound traits.