On The Nature Of Normativity
Interview by Richard Marshall.
[Photo: Gus Ruelas]
‘The real issue is to see how we can integrate our normative and ethical thinking with the insights into the world that we gain from natural science. This is always one of philosophy’s most crucial tasks: to try to see how all of the different kinds of knowledge and understanding that human beings have acquired in different ways – everyday knowledge, historical knowledge, the knowledge that is gained in all of the natural and social sciences, and so on – can be integrated into a single coherent view of the world. In particular, natural science is clearly one of humanity’s most magnificent achievements.‘
‘I don’t believe that understanding a term can ever necessarily involve having false or mistaken views of the world. Understanding a term is like knowing something – it cannot essentially rest on a false belief. So, I don’t “dismiss the norms of witchcraft discourse”. I am engaging in “witchcraft discourse” when I say that there is no such thing as witchcraft. When properly interpreted, the norms of witchcraft discourse – that is, discourse involving the use of a term that has the meaning that ‘witchcraft’ has in English – do not commit those who engage in the discourse to believing that there is any such thing as witchcraft.‘
Ralph Wedgwood is a philosopher who asks questions related to ethics and epistemology. His great-great-great-great uncle was Charles Darwin. Here he discusses the nature of metaethics and why the best-known approaches dissatisfy him, how normativity is to be fitted into a naturalistic framework, his ‘moderate naturalism’, why normativity raises issues of semantics, metaphysics and epistemology, what makes ‘ought’ statements true or false, how we know the real normative components of actuality, how to defend Normative Judgment Internalism, the normativity of rationality, degrees of rational thinking, how he defends his view from various objections, why ‘reasons’ talk is more complicated than many philosophers have assumed, whether rationality is a kind of value and the point of being guided by internal norms. As late summer brings us heatwaves, freshen up on this cool distillation …
3:AM: What made you become a philosopher?
Ralph Wedgwood: I didn’t study philosophy as an undergraduate: I studied Classics and Modern Language (German and Greek) instead. At school, I particularly enjoyed learning foreign languages, including Latin and ancient Greek, and I loved learning about history, literature and culture. However, while I was an undergraduate at Oxford, I had friends who were studying philosophy (for example, Alexander Bird, who is now a Professor of Philosophy at Bristol), and I spent a lot of time talking to them. I also encountered philosophy as part of my studies of German and Greek – in particular, Plato and the Presocratics on the Greek side, and Schiller, Kant, and Nietzsche on the German side. But fundamentally, I just got gripped by the central question of ethics, which Socrates poses so insistently: How should we live? While this is the central question of ethics, in my view answering this question also involves epistemology – since to know how we should live, we need to understand what we should believe, and how we should form and revise our beliefs in response to experience and reflection; and the question as I see it also involves the theory of rational choice or decision—since to know how we should live, we need to understand how we should make choices or decisions, and how we should revise our plans or intentions as we acquire new information over time.
There were many reasons why this question gripped me. I was aware of a number of fundamental disagreements that I had with other people. The mid-80s in the UK was the heyday of Thatcherism, and so I was involved in many political disagreements – for example, some fellow students whom I knew (like the writer Andrew Sullivan) were conservatives, while I was drawn to a more liberal social-democratic view. Some people whom I knew were religious, while I had become a convinced atheist while I was still a teenager. Finally, the fact that I am gay made me acutely aware that some people have moral objections to homosexuality – which also made the issue of moral disagreement highly salient to me.
I became a philosopher because I realized that it was possible to make rigorous systematic progress on understanding Socrates’ great ethical question, by doing philosophy.
3:AM: You’re a leading metaethicist, among other things, in part because of your interest in normativity. Can you briefly sketch what you think metaethics is about and in particular why it’s important? After all, some people may understand the stakes of a practical ethical issue like why racism or inequality is ethically wrong – and they may also get that it’s important to decide the reasons for why we say something is wrong or not – i.e. should we do our moral duty or should we maximize the amount of happiness and so on. But metaethics seems a bit removed from those concerns – so what are the stakes?
RW: For me, metaethics is important above all as a necessary preliminary to studying those ethical issues. In fact, within a few years’ time, I plan to move on from metaethics and to devote myself to a systematic investigation of the most general and fundamental questions of normative ethical theory. I am actually rather dissatisfied with all the best-known approaches within normative ethics – consequentialism, deontology, contractualism, Kantianism, virtue ethics, and so on. Each of these approaches, I believe, gets at part of the truth, but fails rather badly at achieving a convincing overall view of ethics. One reason why these approaches have failed, I believe, is that they have not been informed by a good enough grasp of the metaethical issues. For example, there is supposed to be a big issue in normative ethical theory about whether or not the ‘good’ is prior to the ‘right’ – but the debate has simply not been informed by an adequate understanding of the concepts that can be expressed by ‘good’ and ‘right’, and how exactly they differ from each other. Without an adequate metaethics, normative ethical theory is doomed to fumble around in the dark.
3:AM: One of the general issues facing normativity is how to fit it into a naturalistic framework. Hume put it starkly when he said that you couldn’t get an ought from an is, which I guess helps orientate this issue. I know that naturalism is a rather vague term, so perhaps you could say what you understand by the term and then whether your approach is naturalistic or not?
RW: In my book The Nature of Normativity, I said that my approach was “moderately naturalistic”. (Some of my friends have told me that I should have simply called myself a “non-naturalist”, but I prefer to think of myself as a moderate naturalist.) The real issue is to see how we can integrate our normative and ethical thinking with the insights into the world that we gain from natural science. This is always one of philosophy’s most crucial tasks: to try to see how all of the different kinds of knowledge and understanding that human beings have acquired in different ways – everyday knowledge, historical knowledge, the knowledge that is gained in all of the natural and social sciences, and so on – can be integrated into a single coherent view of the world. In particular, natural science is clearly one of humanity’s most magnificent achievements. In view of the stunning successes of natural science, it has come to seem plausible that natural science enjoys a kind of pre-eminence or primacy among ways of thinking about the world.
It is, however, controversial within philosophy how to understand this “primacy” of natural science. Some philosophers who pride themselves on being “naturalists” hold that natural science should replace all other ways of thinking about the world: for example, epistemology and ethics should be replaced by cognitive science and empirical moral psychology. I believe that this is the wrong approach. Natural science has a kind of metaphysical primacy: the truths that it discovers are metaphysically speaking the most fundamental truths about the world. However, it is a mistake to think that natural science should replace all other ways of thinking about the world, or that the methods of natural science are the best way to understand everything. Ethical and normative questions need to be explored by their own distinctive methods – although as with every other philosophical theory, one of the tests that any answer to these questions has to meet is whether it can be integrated with all of the rest of our knowledge into a single coherent view of the world.
3:AM: In your discussions of normativity you suggest that to understand it properly we need to see it in terms of semantics, metaphysics and epistemology. So beginning with semantics, what are the semantics of normative thought and discourse and why are they important? Is it that you think that by looking at what normative language means –and establishing what you call a ‘conceptual role semantics’ we have to see that the discourse is a cognitive discourse, one where truth and falsehood are very relevant – and in which case versions of expressivism that deny the truth value of normative statements are false?
RW: Yes, establishing that expressivism is false is part of what we learn from studying the semantics of ethical and normative language. It is by no means the whole of what we learn, however! One thing that we can learn concerns the details of the particular concepts that different normative terms can express. For example, I believe that we can explain why the logic of deontic modal terms like ‘ought’ and ‘should’ is so-called “standard deontic logic”. We can also get better insights into the way in which the context in which a term like ‘ought’ or ‘good’ is used determines which of these concepts the term expresses in that context. This can help us to answer such questions as, for example, What do all the different ways of being good have in common, and how do they differ from each other?
We can also learn more general facts about normative and evaluative concepts. In particular, I believe that the best account of what determines the reference of these concepts implies that the properties and relations that these concepts refer to have an intimate connection to the very nature of the mental states that are involved in these concepts’ conceptual role. Specifically, I believe that my account of the semantics supports a version of the idea that is sometimes known as “the normativity of the intentional” – that is, the idea that in giving an account of the kinds of mental states that have “intentional” content, we have to pick out or refer to normative properties and relations.
I now also believe that the semantics of normative language will tell us that the most fundamental normative concepts are concepts of values – concepts of ways in which things can be better or worse than other things. (This is not the view that the ‘good’ is prior to the ‘right’, by the way, since it is my view that rightness and wrongness are themselves values or ways of being good or bad.)
3:AM: Shouldn’t we be cautious about reading too much from the semantics to what’s actual. After all, we’re very comfortable with a semantics that carves the physical world at joints that physicists don’t recognize etc.? Meaning and truth come apart a lot of the time don’t they? To go all Feyerabend on you for a minute, if we can dismiss the norms of witchcraft discourse, why not moral discourse, or rational discourse? Why not anything goes, nothing goes etc.?
RW: This is one view of the relation between semantics and actuality. But I don’t accept this view. I don’t believe that understanding a term can ever necessarily involve having false or mistaken views of the world. Understanding a term is like knowing something – it cannot essentially rest on a false belief. So, I don’t “dismiss the norms of witchcraft discourse”. I am engaging in “witchcraft discourse” when I say that there is no such thing as witchcraft. When properly interpreted, the norms of witchcraft discourse – that is, discourse involving the use of a term that has the meaning that ‘witchcraft’ has in English – do not commit those who engage in the discourse to believing that there is any such thing as witchcraft.
You also raise the issue that normative terms may “carve the physical world at joints that physicists don’t recognize.” I actually think that it is plausible that this is true. But in my view, the fact that physicists don’t recognize these “joints” does not show that these are not real joints all the same. This is part of what I mean by saying that the version of naturalism that I accept is a “moderate naturalism”: in my view, the joints that physicists recognize may be metaphysically fundamental in certain ways, but it doesn’t follow that they are the only joints that exist.
3:AM: So I guess the metaphysical question that needs answering is what makes ought statements true or false? How do you go about answering that question? What properties are there in the world that necessitate normativity? I guess the puzzle for many will be that posed by Hume: if they are is’s then how can they be ought’s?
RW: In my view, there are real normative properties and relations that are exemplified by various items: this is just part of how reality is. Specifically, as I explained above, I now think that the fundamental normative properties and relations are all values – they are all ways in which some things can be better or worse than others. (So, if your acting or thinking in a certain way is necessary for the state of affairs that counts in the relevant sense as the “best available”, then that is enough to make it true that you “ought” to think in that way.) Whenever an item exemplifies one of these values, there is some purely natural feature of the item that in a sense explains why it exemplifies the value. This natural feature will then necessitate the exemplification of the value – and this metaphysical connection between this natural feature and this value is itself just a necessary constitutive truth about the nature of the value itself.
3:AM: So how do we get to know these real normative components of actuality? This I take it is the epistemological issue – and I suppose the big challenge for realists is to explain why if there are real normative facts to be known there is still so much disagreement about norms?
RW: Yes, it is an epistemological issue. On my realist view of normative truth, normative disagreement implies that at least one side in the disagreement has made a mistake and has a false belief. Some normative facts are much easier to know than others. We all know that agonizing pain is normally worse than being comfortable and pain-free; there is virtually no disagreement about that. But we don’t all know that homosexuality is not always wrong – Kant for example thinks that it “debases the human condition below that of an animal, and makes the human being unworthy of humanity”!
I believe that our emotions play a big role in our thinking about values. When we have an emotion, we are normally inclined to form a corresponding evaluative belief (for example, if you find yourself admiring something, this will incline you to believe it to be admirable). However, for our beliefs to be fully justified, we also need to fit them altogether into the most coherent system that we can. When we do this, it becomes clear that some (not all!) of our emotions are misleading and unreliable – they incline us to form evaluative beliefs that on reflection we see cannot be true. (Some social and cultural practices can strengthen these misleading emotions, while others can help us to overcome them.) In addition, it is also sometimes very hard for us to fit our evaluative beliefs together into the most coherent system: we can easily fail to do this very well. These are the two main sources of errors and mistakes about normative and evaluative questions, in my view.
3:AM: It’s part of your approach to defend a version of Normative Judgment Internalism. Can you first say how we ought to understand your approach to this thesis and why you think it very important – and I guess this again raises the issue of why a semantic analysis is going to be enough to establish what it claims. Presumably this is something that is relevant not just to moral thinking but all normative thinking – or even to all rational thinking?
RW: Normative Judgment Internalism (NJI) is the idea that there is an essential (or “internal”) link between normative judgments and practical reasoning or motivation for action. As it is sometimes said, normative judgments are “essentially action-guiding”. In my view, this is what makes normative concepts the special kind of concepts that they are. Some normative judgments (like judgments about what we are justified in believing) guide theoretical reasoning rather than practical reasoning. So the most general formulation is to say that normative concepts are essentially reasoning-guiding.
This kind of internalism has often been misunderstood. It is a thesis about the nature of normative judgments (not about the facts or properties that those judgments are about), and according to a cognitivist approach like mine, the nature of normative judgments is determined by the nature of the concepts that they involve. This is why an account of the nature of these concepts – a semantic account in that sense – has to explain why NJI is true.
Fundamentally, as I see it, NJI is a thesis about the nature of normative concepts. But is has implications for the nature of rational thinking in general. This is because there is something plausible in the idea that thinking explicitly about how one ought to think (or how it is good for one to think) is a way of making explicit to oneself something that is already implicit in simpler kinds of rational thinking. At all events, we have to integrate our account of normative concepts’ conceptual role – that is, our account of how it is rational to use normative concepts – with an account of rational thinking in general.
3:AM: How do you deal with examples where it seems rational not to act on the best reasons available, as in Jake Ross’s Three-envelope problem? Doesn’t this fatally undermine NJI?
RW: No, we just need to give a more careful account of NJI. One issue concerns the individuation of options; another issue concerns the degree of confidence that one has in the normative judgment. Suppose that there are three envelopes, A, B, and C. The fine-grained options are options like Taking A or Taking C. The coarse grained options are options like Taking A or B or Not taking C. We need to focus on judgments about the fine-grained options that are held with maximum confidence, such as judgments like ‘Taking envelope C is best all-things-considered’. If you hold this judgment with maximum confidence, then it is clearly rational to make the corresponding choice and act accordingly.
3:AM: This lets us move away from metaethics and to your very connected but nevertheless distinct interest in the normativity of rationality, the subject of your up and coming book. You argue that all rational metal states have an aim and rationality is the means by which we achieve that aim. So how do you unpack the metaphor of ‘aim’ here?
RW: In the account that I am now advocating, I unpack the metaphor of an ‘aim’ in normative and probabilistic terms. A mental state that achieves its aim is, as I say, “correct”: it is a state that has got things right – while a state that fails to achieve its aim is incorrect or mistaken, a state that has got things wrong. For example, an outright belief is correct if and only if the proposition that is believed is true. This notion of a “correct” mental state is itself a normative notion. Irrationality is connected to correctness probabilistically. The basic idea is that thinking irrationally is bad news about correctness; and the more irrational your thinking is, the worse the news is about how correct your thinking is. This notion of “bad news” is to be cashed out in terms of a certain probability function – intuitively, the probability function that rationally should be guiding your thinking at the relevant time.
3:AM: How does the notion of degrees of ‘correct’ thinking link to notions of degrees of rational thinking?
RW: I want to give a precise account of this idea of irrationality being “bad news” about how correct one’s mental states are. The account that I propose invokes the probabilistic notion of a state of affairs’ “expected value”. In particular, we need the idea of a collection of mental states’ expected incorrectness. There are various possible ways in which things might be – various epistemically possible worlds. First, we measure the degree to which this collection of mental states is correct or incorrect in each of these possible worlds, giving this collection of mental states an incorrectness-score for each of these worlds. Then we take the weighted sum of these incorrectness-scores, weighting each incorrectness-score by the probability of the world. This gives you this collection of mental states’ expected incorrectness.
Then we can say that, quite generally, the most rational thinking minimizes expected incorrectness – and the more irrational your thinking is, the greater the degree of your thinking’s expected incorrectness.
3:AM: You are very aware that not everyone agrees with your approach – they imagine demons might reward irrationality, or people having rational false beliefs about how one ought to think, or come up with objections based on the principle that ought implies can, or that all mental states yearn for is coherence which doesn’t assign any normative importance at all. How do you push these to the kerb?
RW: Part of my response is to return to the semantics of normative concepts. There are many different ways in which things can be truly called “good”. Rational thinking is in one way good, and irrational thinking is in a corresponding way bad. But that is compatible with their being some cases where irrational thinking turns out to be good in some other way (the demon might reward you for thinking irrationally) or rational thinking might be in some other way bad (the demon might penalize you for thinking rationally).
Even if (as I claim) there is a sense in which we “ought” always to be rational, rational false beliefs about how one ought to think will only raise problems if the following three conditions are met: (a) rationality requires us to comply with our own judgments about how we “ought” to think; (b) the sense of ‘ought’ that features in our (rational but false) belief about how we “ought” to think is the same as the sense in which we “ought” always to be rational; and (c) rational false beliefs about how we (in the relevant sense) “ought” to think are possible. I argue that for every relevant sense of ‘ought’, at least one of these three conditions will fail.
It is true that we can’t avoid irrational thinking at will. But that doesn’t show that we can’t avoid irrational thinking at all. I argue that there is a sense of ‘can’ in which we can always think as we are rationally required to think.
Finally, what of the objection that rationality requires nothing more than coherence, which has no real normative importance at all? I accept that in a very broad sense of the term rationality does require nothing more than coherence. (In this sense, we can talk about whether your beliefs cohere with your sensory experiences, episodic memories, and emotions, as well as about whether your beliefs cohere among themselves.) But since irrationality is – as I have explained – bad news about correctness, it seems that it does have a kind of genuine normative significance after all.
3:AM: Why don’t you make a connection between reasons and rationality in your arguments which is what people might expect, given that other philosophers working in this area have?
RW: ‘Reasons’ talk is much more complex than most philosophers working in this area have realized. A number of philosophers hold that there is a single univocal concept of a “normative reason” for an action or an attitude, which is the most basic and primitive of all normative concepts – so that all other normative concepts can be defined or analysed in terms of normative reasons. I argue that this is a dramatic misinterpretation of the semantics of ‘reasons’. There is a huge plethora of different concepts that can be expressed by the term ‘reason’. Moreover, none of these concepts is basic or primitive – they can all ultimately be defined by means of concepts of various kinds of values (along with various explanatory and psychological notions).
Since every concept expressible by using the term ‘reason’ can be defined in other terms, it is unnecessary to formulate these normative debates in terms of “reasons”. Since the term ‘reason’ can express so many different concepts in different contexts, it is positively dangerous to couch these normative debates in terms of “reasons” – we can too easily be led astray if without our noticing, the term ‘reason’ shifts from expressing one concept to expressing another.
3:AM: Is rationality a kind of value?
RW: Yes. Rationality comes in degrees: some mental states and processes of reasoning are more rational, or less irrational, than others. The more rational states and processes are in a way better than the less rational ones. In general, the evaluative concepts that I am interested in are normative (essentially reasoning-guiding) concepts that rank alternative states of affairs (such as the states of affairs of your believing p at a certain time, your suspending judgment about p at that time, your disbelieving p at that time, and so on). The concept of what is “rational” is an evaluative concept of this sort.
3:AM: The take home question is this then: even if we have internal norms, what is the point of being guided by them?
RW: We need to have some norms that can guide us directly – that is, not by means of our doing any further bit of reasoning. These norms that can guide us directly have to be internal norms. But because irrationality is bad news about correctness, being guided by internal norms is a way of pursuing certain external values like truth and making the right choice. You might ask, What is the point of being guided by these external values? But on further reflection, it is clear that this question does not really make sense. It makes no sense to ask: This is true, but why should I believe it? Or, this is the best choice for me to make, but why should I make it?
3:AM: And finally, for those readers here at 3:AM, are there five books other than your own that you could recommend to take us further into your philosophical world?
The deepest influence on my thinking, ever since I began working on philosophy, has been the works of Plato, and especially the Republic. I am fundamentally a Platonist: I believe that the answer to the great question about how to live is ultimately given by certain necessary timeless truths about values – truths that have a mathematical structure that we need to try to understand.
Among contemporary philosophers, I have been influenced by a very wide range of philosophers who have taken different approaches – even though I have significant disagreements with all of them. One philosopher who has made extremely valuable (though also in some ways misguided) contributions to the debate about rationality is John Broome, especially in his book Rationality through Reasoning (Wiley, 2013).
Another philosopher who has influenced my thinking, on normative issues in ethics as well as on more abstract issues about rationality, is Joseph Raz; see for example From Normativity to Responsibility (Oxford, 2011).
On the issue of how to develop a precise formal account of rationality, an excellent recent book is Richard Pettigrew, Accuracy and the Laws of Credence (Oxford, 2016).
Finally, on the semantic issues that are so important for avoiding methodological pitfalls, I recommend a new collection of essays edited by Matthew Chrisman and Nate Charlow, Deontic Modality (Oxford, 2016).
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
Buy his book here to keep him biding!
First published in 3:AM Magazine: Friday, August 26th, 2016.