The Is/Ought Gap: It’s real but doesn’t pose a problem for “scientific moralists”

In the past two weeks I’ve stumbled upon blog entries that argued for a “scientific morality”, and in doing so challenged the “is/ought thesis”, which seems to be a rite of passage to be a naturalist these days. Unfortunately, the writers misunderstand the “is/ought” thesis, and so their arguments against it fail (but both entries are interesting and worth reading, in no particular order, here and here). Both writers seem to think that the “is/ought” thesis is false because it entails that evaluative statements are not factual statements, and as statements about value are facts when they are true, the thesis is false. Indeed, what makes a statement a fact is that it is true, and so evaluative statements and descriptive statements are true if they are propositions that correspond to the world. In order for their arguments against “is/ought” to be correct it would have to be the case that the “is/ought” thesis entails that the difference between evaluative statements and descriptive statements is that evaluative statements do not describe the world, and so cannot correspond to the world in a way that would make them truth-apt propositions, and thus possible of being facts when true. But the “is/ought” thesis does not entail this at all. All that the thesis requires is that the difference between evaluative statements and descriptive statements is that evaluative statements contain evaluative terms, while descriptive statements contain no evaluative terms. It would take further argument to show that a statement containing evaluative terms cannot be truth-apt.

Today I want to contribute to this discussion of the “is/ought” gap, with the hopes that I will convince some of the truth of said gap. A further point will be to suggest that accepting the “is/ought” gap does not rule out a scientific morality.

The “Is/Ought” Gap:

The “gap” between the “is” and the “ought” is the gap in a deductive argument where there must be a leap from purely descriptive premises to an evaluative conclusion. Consider this argument:

P1. Factory-raised livestock endure much pain and suffering through the course of their lives.
P2. To eat factory raised livestock is to financially support factories that cause livestock pain and suffering through the course of their lives.
C. Therefore, it is morally wrong to eat factory-raised livestock.

This argument is not valid because the truth of the premises does not guarantee the truth of conclusion, as it seems possible that given the truth of these premises it might be false that it is wrong to eat factory-raised livestock (as evidence consider that many omnivores grant the truth of the premises but deny the conclusion). What we need is a “bridging premise” that makes the conclusion necessarily follow from the other premises. The needed premise would be “it is morally wrong to financially support factories that cause livestock pain and suffering through the course of their lives”, as we can see clearly that if eating factory raised livestock financially supports pain and suffering, and financially supporting pain and suffering is wrong, then eating factory raised livestock is wrong. But this of course is to introduce an evaluative premise into the body of argument, and so, here at least, we see that an evaluative conclusion doesn’t follow from descriptive premises unless there is an evaluative premise in the argument.

The defender of the “is/ought” thesis is prepared to press this point against all such arguments, showing those arguments with only descriptive premises that seem valid to only be valid because of an implicitly evaluative premise hidden in the argument. Of course, this method of argumentation would only be to show on a case by case basis that it is false that “is/ought” gap is false. Nevertheless, there is good reason to think the “is/ought” gap is real, because the “is/ought” gap is a moral manifestation of a general principle of logic that goes, “one cannot validly deduce a conclusion that applies a predicate P, from premises that nowhere contain that predicate. This is nothing special about evaluative predicates; it applies to any predicate whatsoever” (Michael Huemer, Ethical Intuitionism, page 74). Thus, there is a powerful argument for why the gap is real, and good evidence to support the generalization that no argument with all descriptive premises can lead to a evaluative conclusion without an implicit evaluative bridging premise.

The Consequences of Embracing the “Is/Ought” Gap:

Contrary to popular belief the consequences of accepting the “is/ought” gap are not very serious. One response to the “is/ought” gap is to note that, as it does seem that we have moral knowledge about substantive moral positions, and to do so we must have moral knowledge that is not deducible from purely descriptive premises (that is, after all, the nature of the gap), there must be some bedrock, foundational moral knowledge that is a priori. In the example given above, that a priori moral knowledge might be, “it is morally wrong to cause pain and suffering”. Indeed, if there is any a priori moral knowledge, it is likely to be about the wrongness of causing pain, as we can see on reflecting on the nature of pain as a concept that there is an impartial to-be-avoided-ness built into the concept, thus amounting to a priori knowledge seen through what is called ethical intuition (see my writings on Robert Audi or ethical intuitionism in general). I favor this response.

The other route is to accept the “is/ought” gap, and with it that evaluative conclusions cannot be deduced from purely descriptive premises, but posit that cogent inferences can be made from descriptive premises to evaluative conclusions. The Cornell realists (the original proponents of scientifically reputable moral realism in the 20th century) embrace this conclusion with Nicholas Sturgeon noting in “Ethical Intuitionism and Ethical Naturalism” that we can breach the gap with an inference to the best explanation, a method grounded in an empiricist rationale and thus being entirely consistent with common methodologies of scientific inquiry, and thereby the possibility of a scientific morality. David Brink, another influential Cornell realist writes that, “thus, the is/ought thesis leaves open the possibility […] of good nondeductive inferences between moral and nonmoral statements with, and perhaps without, the benefit of bridge premises. […] I assume that, whatever its exact specification, inference to the best explanation is a legitimate nondeductive inference pattern” (David Brink, Moral Realism and the Foundations of Ethics, page 169). The point is that we can have moral knowledge and be scientifically minded moral realists and still embrace the is/ought gap, which we have good reason to accept.

Conclusion:

The “is/ought” gap is quite real, but scientifically minded moral realists don’t have much to fear from it. I’m not sure why they feel the need to resist it. It seems they purposefully conflate it with a metaphysical thesis concerning reductivism, and thus, the attack on the “is/ought” thesis is in the absence of successful arguments for reductivism, which seems to be the other component of “scientific realism”.

Advertisements

, , , , , , , , , , , , , , , , , , , ,

  1. #1 by skyscraperbexten on October 11, 2014 - 6:31 pm

    Thank you for your time, ausomeawestin.

    I don’t “purposefully conflate” any issue. I’m not trying to deceive anybody. I’m not talking about the is-ought problem because I want to hide “the absence of successful arguments” for my position, whatever you think it is. I’m honestly trying to understand the best criticisms of my position.
    It also seems that you think I’m sneaking in “evaluative premises” into my arguments. I just don’t see it that way. Take this argument form:
    If you want x to happen, then you ought to do y.
    You want x to happen.
    Therefore, you ought to do y.
    The example argument that you gave contained a hidden premise. Mine didn’t. They’re not equivalent.
    When I talk about morality, I define it as “what one ought to do more than anything else.” So it takes the argument form:
    If you want X more than anything else, then you ought to do Y more than anything else.
    You want X more than anything else.
    Therefore, you ought to do Y more than anything else.
    If somebody has a different definition of morality, then it won’t be what one ought to do above all else.
    My account of morality depends on two things: the existence of agents with different goals and different methods of achieving those goals. What I’m “evaluating” is the effectiveness of a certain method with regard to a certain goal. To me, an “evaluative” fact is just another descriptive fact.

    • #2 by ausomeawestin on October 11, 2014 - 7:25 pm

      And thank you for responding to my post.

      I understand that the argument form you give is different than the ones I’ve been discussing. But insofar as yours is different from those arguments it is irrelevant to the is/ought problem. The is/ought thesis is a very specific problem about deductive arguments, and your response to it is not adequate if it doesn’t address the original question of how an argument can have purely descriptive premises and an evaluative conclusion. The first premise in your argument states the logical connection between a descriptive premise and an evaluative conclusion that a response to the is/ought problem is out to provide — it’s question-begging. My point is that this is not up for debate; the is/ought problem is a fixed problem about arguments of the form p1. x, p2. y, c. so z., that philosophers have been discussing for some time, so you can’t just ignore the fact that the problem is only about this form of argument, and claim a different form of argument, as it is no solution to say p1. if x then z, p2. x, c. so z. when that is not what is being discussed. You’re right that we can state moral arguments that way, but that is besides the point of the is/ought problem.

      What is more, your argument does not count against the is/ought argument because it has a moral term in a premise in the body of the argument. Saying that “ought” is a descriptive term, as I know you want to do, is fallacious; yes, it is term that refers to/describes a property, but it refers to an evaluative property, a normative property that is about reasons to do something, so it is a term that refers to an normative property and thus, for philosophers engaged in this debate, is a moral term, and not a descriptive term. You cannot walk in the door and change the definitions of words that have been being used in this debate.

      On this subject, in your last sentence you say “To me, an “evaluative” fact is just another descriptive fact.” This is where I get the idea that you want to reduce moral facts to non-moral facts and falsely attack the is/ought thesis on the grounds that it goes against reductivism (which it doesn’t). Evaluative facts cannot be just descriptive facts but they are still facts, and this if the evaluative property said in the statement accurately corresponds to a property that an action has (correspondence theory of truth), because evaluative statements can describe an action or event with evaluative terms. That is the difference between evaluative statements and descriptive statements, the former contains evaluative terms while the other does not, both describe states of affairs. Evaluative facts cannot be just descriptive facts because descriptive facts contain no evaluative terms, in that they use no terms that refer to normative properties. Evaluative statements describe, but they use terms that refer to normative properties. You’re reading too much into the idea that our dichotomy is between “descriptive” and “evaluative” facts, falsely inferring from this that evaluations cannot be descriptive, but they can be. Please revisit my first paragraph of the original post above. It seems to me you want to say that descriptive statements and evaluative statements have different functions, but it need not be that way, and many philosophers engaged in the is/ought debate think that both types of statements describe, the difference is just for organizing our thinking, where we say a statement is evaluative if it contains an evaluative term and thus refers to a normative property, or a statement is descriptive if it contains no evaluative terms, and thus, does not refer to any normative properties. The point of all this is that evaluative statements can be descriptive, and thus truth-apt, such that when true they are facts, such that we don’t need to say that evaluative statements must actually be descriptive statements if they true facts, as you seem to assume.

    • #3 by SelfAwarePatterns on October 11, 2014 - 7:39 pm

      Hi skyscraperbexten,
      Your argument seems valid, but it doesn’t seem like a moral one to me. If my version of X is to spread my genes maximally or to cause the maximum amount of suffering possible, then will the resulting what-I-ought-to-do be moral?

      • #4 by skyscraperbexten on October 11, 2014 - 8:15 pm

        If inflicting the most suffering was what I wanted most, then I would say yes. But I actually think I want welsh corgis more;).

        An important idea that I *ought* to have mentioned was the “rational and sufficiently informed” condition. If I was rational and had enough information to know what I wanted most, then that would be what I would be striving for.

        I’m currently researching sociopaths in my spare time. Do sociopaths actually enjoy being sociopaths knowing the full implications of that condition? If they do, that might count against my position. If, however, they don’t want to be sociopaths and wanted to change, then they’ll won’t want to inflict suffering.

        Like all of science, morality would depend on the facts, regardless of what we want them to be.

      • #5 by SelfAwarePatterns on October 11, 2014 - 11:28 pm

        The problem is that two people can be identically informed and still have different preferences due to innate genetic tendencies. For example, knowing the same information about the threat of terrorism, some people will prefer stricter security policies even if it curtails their freedoms, while others will insist the trade-off isn’t worth it. Neither knows what terrorist attacks actually will or will not occur or if they themselves will be affected. So who, logically and scientifically, has the right priorities between freedom and security?

  2. #6 by SelfAwarePatterns on October 11, 2014 - 7:19 pm

    Well said. Science cannot determine values, but it can inform those values in a major way. Science can study what the effects of corporal punishment might be, but our response to knowing those effects will be based on our values. So science has a role to play in morality, but it’s important to understand what its limitations in that role are.

    As you know, I see morality arising from a combination of instincts and cultural learning. People’s instincts and cultural background will vary, but there is enough of an overlap among us for society to work. This view makes me something of a semi-realist.

    • #7 by ausomeawestin on October 11, 2014 - 8:11 pm

      Thanks, Sap! I’m definitely in full agreement with your thoughts in the first paragraph. Though I favor intuitionism as giving us foundational moral beliefs, I definitely think coherence among higher-order beliefs is to be aimed for, as I see the sciences as playing a role in informing and expanding these beliefs, as you say.

      In the past I’ve resisted the sorts of ideas mentioned in your second paragraph. But I’ve been considering the idea that instincts might be the most common source of a priori moral knowledge. I had been worried in the past that that would indicate morality’s dependency on human cognition, a notion I was uncomfortable with, thinking it would make morality not mind-independent in a way that would not be objective. I’ve come to see that morality could still be objective and not entirely mind-independent, and in fact, I’ve come to think that moral truths are crucially mind-dependent, in being realistic about moral properties, as they are conceptual truths that are seen in reflecting on natural properties. Furthermore, I’ve come to think that morality could be entirely mind independent in the way I don’t believe in but it still be the case that we come to have moral knowledge through instincts. The idea would be Kantian in inspiration, epistemologically that is. Our cognitive capacities shape and structure our experiences in ways that present reasons for certain actions to be done, given the stimuli we receive from the world, which I think is one way of understanding instincts, preservation or otherwise. These same capacities could structure our experience of natural events in such a way that there is a normative reason for an action to be done for the benefit of another, and in such an instance, our instincts have generated moral belief, which would be enough to say it is a priori. Whether we could say that belief is true could depend on other internally accessible evidence, or even the sort of data gained from scientific studies that we discussed above. The point is that moral beliefs generated from instincts might be reliable, and this epistemological question, as to whether such beliefs are justified, is separate from the question of whether such beliefs are true, which would require further metaphysical arguments.

      All of this goes to say that I find myself more and more open to your ideas. Thanks for sharing them with me!

      • #8 by SelfAwarePatterns on October 11, 2014 - 11:16 pm

        My pleasure. I learn a lot from you about moral and philosophical frameworks and terminology, and you occasionally force me to rethink or at least substantiate my positions, so it’s a rewarding exchange!

        One point I would make. When talking about mind independence, I think a distinction should be made between being dependent on our conscious mind versus being dependent on evolved instincts. The former, it seems to me, implies cultural relativism, while the latter allows for some degree of naturalistic realism. (Never one to make things simple, I think reality is a mix of the two 🙂 )

      • #9 by ausomeawestin on October 12, 2014 - 4:49 pm

        That’s a very interesting distinction to make. I wonder, though, if morality could depend on the conscious mind without implying cultural relativism, being as the conscious mind could see conceptual connections between natural properties and normative properties, or properties that create reasons for action, which would allow for objective and not relative facts. The relations could only be available to a conscious mind, in such a way that the relations are not out in the world, but this would not imply relativism unless we said that natural properties could conceptually entail different normative properties depending on the culture/society in which the observer is located. So I would resist the point that dependence on conscious minds implies relativism.

      • #10 by SelfAwarePatterns on October 12, 2014 - 5:21 pm

        Interesting. I had two reactions to your point. My first was that since those relations exist independent of our conscious mind, they wouldn’t be dependent on it, at least not completely. My second reaction was an epiphany that all moral constructs are dependent on our conscious minds, since they wouldn’t exist otherwise.

        I think the important distinction is not whether conscious minds are required, but are they sufficient? That said, my views varied while typing this, so color me undecided 🙂

      • #11 by ausomeawestin on October 12, 2014 - 6:00 pm

        Ah yes I see your thinking. The sort of mind dependence I had in mind was the first sort, where minds are needed to grasp the relations, but where the relations are not determined by minds. If the latter were true then some sort of subjectivism would be true, possibly cultural relativism, as you say. I think these relations could be grasped by alien minds, in a certain way of accommodating multiple realizability for moral properties, such that normative properties could exist in relation to natural properties even if human beings did not exist.

        The distinction between whether conscious minds are necessary or sufficient for morality is a very interesting one. My first reaction is that, without specifying further what a conscious mind includes in the background of its abilities, it is not. We need to have some experiences and learn something about sound reasoning before a legitimate sense of morality can be said to be had.

  3. #12 by Larry on October 12, 2014 - 7:59 pm

    Thanks for a very interesting post.

    Regarding your statement that “It seems they purposefully conflate [the “is/ought” gap] with a metaphysical thesis concerning reductivism”. You defined “reductivism” in an earlier post this way: “a reductive account of normativity would … provide necessary and sufficient conditions for a normative property in non-normative terms”.

    I’ve always thought the existence of the is/ought gap serves as an argument against this kind of reductive account of normativity (and a better argument against such accounts than Moore’s “open question” argument). In other words, it’s a mistake to think that normative language can be analyzed in terms of or reduced to or inferred from descriptive language, as ethical naturalists have often tried to do, given the existence of the is/ought gap. At some point, as Hume argued, a normative language or assumption will have to come into play in order to allow the transition to a normative conclusion. I don’t think talk about ethical properties supervening on natural properties avoids the problem.

    On the topic of moral knowledge, it’s true that IF we do have moral knowledge, it must come from somewhere, but concluding that it must be a priori or from a faculty of ethical intuition doesn’t seem to clarify matters. We might just as well say that most of us have very strong ethical intuitions, which we feel so strongly about that we can’t imagine we’re wrong (we just know we’re right).

    Similarly, on the page from Brink’s book that you cite, he says: “Given only nonmoral background assumptions, for example, that moral realism is true [I guess that’s a metaphysical assumption] and the appraisers in question are fully rational and fully informed, the moral fact that [torture] is wrong may provide the best explanation of the nonmoral fact that appraisers unanimously agree that [a given example of torture] is wrong”. He goes on to say that one may question the background assumptions, in particular, that moral realism is true (a claim that he has argued for). Another explanation for the unanimity, of course, is that the appraisers are human beings who grew up in the same culture.

    I think that “scientifically minded moral realists” can agree, as you say, that the is/ought gap exists, but making the case for moral realism would be much easier if it didn’t.

    • #13 by ausomeawestin on October 13, 2014 - 8:45 pm

      Thank you for reading the piece, and for your insightful comments on it. I hope that you will not object to me turning your comments and my responses to them into a new post; I think your comments are remarkably valuable in their incisiveness.

  4. #14 by rung2diotimasladder on October 14, 2014 - 10:10 am

    Excellent post! My only experience with this argument is from reading Sam Harris’ book, The Moral Landscape, which I wrote about here: http://theleatherlibraryblog.com/the-fatal-flaw-in-sam-harris-the-moral-landscape-how-science-can-determine-human-values/

    Not sure how I feel about the is/ought distinction. I tend to think there is one, but there’s always going to be someone to call any example of irreducible “ought” meaningless or not true. It gets frustrating.

    • #15 by ausomeawestin on October 16, 2014 - 11:17 pm

      Thanks for the kind words! I had actually read that post of yours a while back (I enjoy the leather library), and very much liked your entry on Harris’ book, will have to revisit it (I don’t care for Harris).

      I actually favor an anti-reductivist approach to normative properties, so it’s been funny responding to comments/objections here to defend the view that a thesis I think is true (the is/ought gap) doesn’t imply the falsity of a thesis (reductionism) that I nevertheless think to be false for other reasons.

      • #16 by rung2diotimasladder on October 17, 2014 - 1:52 pm

        LOL! I love it. Sometimes such subtleties are hard for people to grasp. I often get this with politics. I’ll argue against some reason for believing in an X that’s typically “liberal” and people will assume a) I don’t buy into the X for other reasons and b) I’m a Republican.

  5. #17 by jmeqvist on October 14, 2014 - 9:06 pm

    Awesome post! I agree with your account, but as I was reading through it I began to think about MacIntyre’s critique of the ought/is distinction.

    I assume you are familiar with it, but I will briefly summarize it. MacIntyre says that the ought/is distinction only makes sense against the backdrop of moral philosophy after teleology had been purged from its vocabulary. He makes the interesting and compelling point that functional concepts allow one to derive an ought from is. For example, by the fact that a watch is uncomfortably heavy or does not keep accurate track of time, we can justifiably say that this watch is bad, The idea being that the notion of a watch is a functional concept, and so we cannot really separate an evaluative and descriptive premise, as the two are inextricably fused when we are dealing with functional concepts. What would you say to this argument? Because if MacIntyre is right it seems we can derive oughts from an is in the case of teleological or functional concepts.

    It should be noted I do recognize that even if this argument is valid, it stands only if teleology is justifiable, and teleology is extremely hard to defend in any robust form.

    • #18 by ausomeawestin on October 18, 2014 - 3:41 pm

      Thanks for the kind words and the fascinating comment; I must admit that I haven’t read MacIntyre, so thanks for sharing this interesting objection to the is/ought gap.

      Admittedly, I don’t find the objection very convincing, because I think insofar as the premises lead to the conclusion, there is a hidden premise/assumption doing all the work in bridging the gap. It seems the argument,

      p1. My watch doesn’t tell time.
      c. Therefore, it is a bad watch.

      is only valid if we make explicit the hidden premise/assumption, p2 “A watch should tell time”. I have a watch that doesn’t tell time, but it is not a bad watch, because I am not using it to tell time. I’m using it as a prop for Halloween – I’m going as Bob McDonnell, the corrupt Virginian governor who received a Rolex, among other gifts, from someone trying to gain influence. One might argue that the watch I’m speaking of has a different functional role, and so doesn’t serve as a counterexample, but that doesn’t seem right, because the purpose of the faux-Rolex is to represent a watch, and it can perform this purpose whether or not it actually does tell time, so I think the fair conclusion to make is that “watch” can have a purely descriptive meaning. In that sense ‘watch’ still has a meaning apart from some sort of functional role, and this role is descriptive and meant merely to label an object. To assume the functional role of an object is to bring an evaluative consideration into the argument, and consistent with there being a gap, to introduce the evaluative premise needed to get from descriptive premises to an evaluative conclusion.

      I think this point, that any argument that leads from seemingly descriptive premises to an evaluative conclusion has a hidden evaluative premise or assumption in the argument, can be pressed against all such examples. The use of teleological or functionally-loaded concepts is an attempt to sneak in evaluative terms disguised as descriptive terms, but if they are really and truly teleological or functional concepts then they are evaluative assumptions, and the fact that having them in an argument can lead to an evaluative conclusion is not a counterexample to the existence of the is/ought gap.

      • #19 by jmeqvist on October 20, 2014 - 8:37 am

        Thanks for the insightful response. I have the same issues with MacIntyre’s objection to the is/ought gap that you do. It seems to me that as long as the evaluative/descriptive distinction holds the is/ought gap will hold.

  1. The Is/Ought Gap pt. II: On Its Implications for Reductionism and Intuitionism | ausomeawestin
  2. Is/Ought, and Popping the Hood on Argument | Amusing Nonsense

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: