Follow up to: Too Many Moralities

After putting it aside for awhile, I think it’s time to take a bit more of a trek on my quest to understand normativity, which is basically about figuring out what words like good/bad, right/wrong, better/worse, ought/ought not, should/shouldn’t, etc. mean.

I borrow the analysis of philosopher Stephen Finlay in proposing end-relational theory where these words are all taken to express how something relates to an end — a cup isn’t good, it is good for drinking and saying “I ought to use this cup instead of that one” means “Using this cup instead of that one is more likely to lead to me enjoying my drink”.

However, the end in mind isn’t always clear, and statements like “you ought not murder!” certainly don’t sound like they’re just about satisfying a goal. This is because ends are often implied only by context, removed for rhetorical effect, and stated in the form of a demand.

Once we recognize this, we realize that there is not just one thing we ought to do. We benefit by being able to cross the is-ought gap, but run into an apparent difficulty when we recognize that instead of one morality, there are many moralities. In this essay, I want to dig us a bit further out of this apparent ditch.

 

The Ought Simpliciter

A similar approach to Finlay’s is also explored in Judith Jarvis Thompson’s book Normativity. Thompson starts also by arguing that something is “good” if it meets a certain standard of “good-making” and agrees that whether someone is morally good (good at meeting moral standards) has nothing to do with what they desire or what they have practical reason to do; instead they always have a moral reason to meet moral standards and morally ought to meet those standards, whether they want to or not.

However, where Thompson diverges is on the concept of “ought” simpliciter. Thompson argues that even if someone wants to kill someone by poisoning their coffee, stands to gain a lot by killing the person, and pragmatically ought to poison their coffee, none of this matters because it just isn’t true that they ought to poison coffee. People just plain ought not kill people.

This notion of a “just plain ought” is what is meant by “ought simpliciter” — a straight up normative ought that transcends all ends and goals, and just is what we ought to do. However, I think that if you’re following along the logic of what “ought” means, then the idea of an ought simpliciter doesn’t make any sense; it’s a vacuous and meaningless term.

 

Thompson suggests that one just plain ought not kill people. But what is meant by this? On Finlay’s analysis, “we ought not kill people” reduces to “killing people is likely to prevent us from satisfying {goal}”. However, if we leave “ought” in a floating simpliciter form, we are leaving the {goal} unspecified and thus should envision some sort of ERROR 404 GOAL NOT FOUND whenever we try to wrap our head around it.

I think it quite obvious that what Thompson has in mind for “we ought not kill people” is “we morally ought not kill people”, which defines {goal} as “contextually implied moral standard”, which allows us to swap “ought” with “violates moral standard” and lets us continue to process the question.

 

 

Why Not Just Define “Ought” Simpliciter As “Moral Ought”?

Once we accept that saying someone ought to do something must require an associated goal or standard (even if implied and not directly stated) before we can understand what is being said, it becomes clear that there is no such thing as ought simpliciter in a strictly logical sense. But saying “you ought not kill” is much punchier than saying “you morally ought not kill”; so why not get the benefit of the simpliciter rhetoric and just agree that when we say “you ought not kill” we mean “you morally ought not kill”?

Honestly, I think this is fine for the most part. I think it worthwhile to make use of rhetoric to promote our moral aims, and thus when we’re out in the field so-to-speak persuading people to do what they morally ought to do according to our preferred moral standard, but when we’re trying to do careful analysis, we don’t want to fool ourselves in thinking that our preferred moral standards are on any stronger ground than they actually are.

What I mean by this is that declaring some goal to be what we mean by an unspecified ought simpliciter is to give that gaol a special status of being more important than other goals from a meta-ethical point of view, which is a privilege that cannot be justified from this meta-ethical view. There just isn’t a meta-ethical basis to declare one goal better than another; goals are only more or less important in the context of desires and/or other goals.

Taking the goal that we want to avoid confusing ourselves when doing analytical meta-ethics, we ought not have an ought simpliciter.

 

 

Ought, All Things Considered

Another consideration similar to ought simpliciter is that of “ought all things considered”. What we “ought, all things considered, to do” is that which we ought to do when reflecting upon all possible goals. For example, given that I know I morally ought not poison her coffee but I pragmatically ought to, what ought I, all things considered, do?”

However, the idea of an “all things considered” ought has a problem different than ought simpliciter, but just as bad. Recall again my definition of ought — “Ought {goal} X = X best satisfies {goal}”. This would mean that we all-things-considered ought do whatever best satisfies all goals.

Ought simpliciter has the problem of referring to no goals, whereas ought all-things-considered has the problem of referring to all goals. And given how many goals are self-contradictory, this won’t work out. How can we simultaneously satisfy a goal to poison someones coffee and a goal to ensure that coffee remains free from poison?

 

While this is the downfall of “all-things-considered” taken literally and analytically, this does not mean that the concept is dead in spirit. Instead of considering all possible goals, we could limit our consideration to all possible goals we care about, or a consideration of the second-order goals needed to accomplish a first-order goal.

Return to the tension between wanting to poison someone’s coffee for a pragmatic payoff (maybe you’ll get a million dollars in inheritance) while violating a moral standard. Perhaps here you could realize: “I care about both what’s good for me pragmatically and not violating my moral standard. Given this tension, what ought I, all-things-considered do in order to get what I most desire?”.

This is really a question about one goal — the maximization of a personal desire set. However, there are two subgoals in this desire set at play — the desire to be moral and the desire to gain the monetary payoff. Thus, there is a question of considering and weighing multiple goals against each other, but there’s a way that we avoid conflicts between goals — by comparing them to a main goal.

I think as long as we’re careful about how we employ “all things considered”, we can use it to capture this important concept. Though I offer that as a speculative conclusion, ready to discard “all things considered” language if its more trouble than its worth.

 

 

Conclusion

Oughts are, by the very definition I have given them, stuck relative to a target goal or end, thus asking what we ought to do without specifying a goal will only result in ERROR 404: GOAL NOT FOUND and the inability to process the question at hand. But why do people keep insisting there is an ought simpliciter? I suspect two reasons: (1) we’ve grown so accustomed to masking the goal in question that its natural to assume ought can exist without one and (2) we prioritize questions of moral normativity so highly over that of other normative questions that we think of normativity solely in terms of our preferred morality, and thus assume “ought” must be linguistically reserved for our morality.

But keeping with an ought simplicter, while rhetorically effective, is not analytically effective. Thus, taking the goal of not confusing ourselves when we do analytical work, I recommend it not be done. Likewise, I recommend caution when working with an ought all-things-considered, since you are still considering goals in the context of a first order goal that must be specified.

 

If I were to provide one lesson about meta-ethics, it would be this: in all cases, you need to specify a goal in order to make sense of “ought”, one way or another. That’s the key aspect of the end-relational view, and I think once this is recognized as true, a lot of meta-ethics instantly falls into place and makes sense. At least, that’s been my experience.

Now that I’ve argued for this view of meta-ethics, I’ll finish up the puzzle I opened with “The Meaning of Morality” and try to answer the questions I laid out there. After that’s done, the meta-ethics will all be in place, and anything additional from me should just be clarifications, revisions, or answers to objections.