The Golden Rule: A Defence

ABSTRACT

According to the so-called golden rule, we ought to treat others as we want to be treated by them. This rule, in one form or another, is part of every major religion, and it has been accepted by many philosophers with various ethical views. However, if the literal golden rule is interpreted as an absolute rule, it is problematic. In this paper, I introduce a new version of this famous principle that is similar to various classical definitions. According to this variant, the rule can be formulated in the following way: If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and you have a good will, then you ought to do $H$ to x. I show how this version can be derived from a small set of highly plausible premises, and I defend it against some of the most interesting and/or common objections against the golden rule in the literature. I conclude that we have good reason to believe in this rendition of the principle.

KEYWORDS:

• The golden rule
• the principle of karma (cosmic justice)
• moral principles
• conditional norms
• practical rationality

1. Introduction

The so-called golden rule (GR) is one of the most famous moral principles,¹ having been a part of our ethical discussions for more than two thousand years,² and, in one form or another, it is included in every major religion.³ Many religious and non-religious thinkers with different backgrounds and with different normative and metaethical views have accepted the golden rule.⁴ There are many different interpretations of this influential principle, some more plausible than others.⁵ Here is a literal version:

LGR If you want $x$ to do $H$ to you, then you ought to do $H$ to $x$. $\mathrm{WyHxy}\to \mathrm{OHyx}$.⁶

If LGR is interpreted as a rule of thumb, it seems reasonable, but if it is interpreted as a literally true (necessary) principle, it is problematic. In Section 2, I will mention some problems with this reading and with some other similar principles. In this paper, I will explore a new rendition of the golden rule that is similar to LGR and various other classical versions. According to this version, the rule can be formulated in the following way:

SGGR If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and you have a good will, then you ought to do $H$ to $x$. $\left(\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\wedge \mathrm{Gy}\right)\to \mathrm{OHyx}.$⁷

SGGR should be distinguished from the following variants of the golden rule:

GGR If you want $x$ to do $H$ to you and you have a good will, then you ought to do $H$ to $x$. $\left(\mathrm{WyHxy}\wedge \mathrm{Gy}\right)\to \mathrm{OHyx}$.

SGR If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, then you ought to do $H$ to $x$. $\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\to \mathrm{OHyx}$.⁸

SGGR differs from LGR in two ways: it includes a ‘same-situation clause’ and a ‘good will clause’. The idea that the golden rule ‘needs’ a ‘same-situation clause’ is old and seems to go back at least to Francis of Assisi.⁹ In Section 2, we will see how this clause helps us avoid certain problems with LGR. Note that the attitude takes wide scope in this condition. It says, ‘if you want it to be the case that if $x$ were in your situation and you were in $x$’s situation, then $x$ would do $H$ to you’, not ‘if $x$ were in your situation and you were in $x$’s situation, then you would want $x$ to do $H$ to you’. The narrow scope reading generates another version of the golden rule, namely the following:

PGR If it is true that if $x$ were in your situation and you were in $x$’s situation then you would want $x$ to do $H$ to you, then you ought to do $H$ to $x$. $\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{WyHxy}\right)\to \mathrm{OHyx}$.¹⁰

In our symbolisation of SGGR, we use ‘$S^{y}x$’ to say that $x$ is in your situation and ‘$S^{x}y$’ to say that you are in $x$’s situation. But what does it mean to be in someone’s situation? We can think of ‘$S^x$’ as an abbreviation of a conjunction of all properties that $x$ in fact has (in a possible world at a particular time) that we are interested in. According to the strongest interpretation of this claim, this conjunction includes all of $x$’s (actual) properties (including various relational properties). But there are weaker interpretations where we focus on only some properties. To avoid various problems, we will only include ‘universal’ properties in ‘$S^y$’ and ‘$S^x$’ (someone’s situation).¹¹ A universal property is a property that can be expressed without proper names (‘Plato’, etc.) and ‘indexical’ words (‘I’, ‘this’, ‘that’, etc.).

To illustrate what it means to be in someone’s situation, consider the following example. Suppose that Peter is a teacher and has children and that Anna is a doctor and does not have children. Then in a reversed situation, in a possible world where Peter is in Anna’s situation and Anna is in Peter’s situation, Anna is a teacher and has children and Peter is a doctor and does not have children. But Peter still has the property of being identical to Peter in this situation and Anna still has the property of being identical to Anna in this possible world.¹²

SGGR also includes a ‘good will clause’. The idea that the golden rule seems to ‘need’ some condition of this kind is also old.¹³ In Section 2, we will see how this clause helps us avoid certain problems with LGR and SGR. What does it mean to have a good will? We shall say that an individual has a good will if and only if everything this person wants ought to be.¹⁴ In fact, in our derivation of SGGR we only need the weaker claim that an individual has a good will only if everything she wants ought to be. Note that we do not require the ‘converse’ proposition. An individual that has a good will does not necessarily want everything that ought to be the case. It is probably very difficult to have a good will, but it is not impossible.¹⁵

We seem to need a ‘same-situation clause’ and a ‘good will clause’ in the golden rule to avoid certain problems, but we also need them for our derivation of SGGR in Section 3 to go through. So, in a sense, these clauses are not just some ad hoc conditions that we introduce to avoid some problems with the literal golden rule. They are ‘essential’ to this version.

We assume that ‘$H$’ in SGGR stands for simple, dyadic action predicates, like ‘help’, ‘tell the truth to’, ‘treat with respect’, etc., and negations of such predicates, for example, ‘not kill’, ‘not lie to’, ‘not rape’, ‘not steal from’, etc. Hence, the following proposition is an instance of SGGR: if you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would help you, and you have a good will, then you ought to help $x$. This also means that the following version of the so-called ‘silver rule’ follows from SGGR:

SGSR If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would not do $H$ to you, and you have a good will, then you ought not to do $H$ to $x$. $\left(\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{\neg Hxy}\right)\wedge \mathrm{Gy}\right)\to O\mathrm{\neg Hyx}$.

Here are some instances of this principle: if you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would not kill you, and you have a good will, then you ought not to kill $x$; if you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would not lie to you, and you have a good will, then you ought not to lie to $x$; and if you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would not steal from you, and you have a good will, then you ought not to steal from $x$.

In the next section (Section 2), I will mention some potential problems with the golden rule, and I will show how our version of this principle, SGGR, can be used to avoid these problems. In Section 3, I will prove that SGGR can be derived from a small set of highly plausible premises. Section 4 is a brief summary of the paper where I conclude that we have good reason to believe in the version of the golden rule, SGGR, that is explored in this paper.

2. Objections to the Golden Rule

In this section, I will mention some potential problems with the golden rule. There are many arguments against this principle in the literature; so many, that I can only discuss a small number of them in this paper. I will focus on three kinds of objections that seem the most interesting and/or common to me. Furthermore, I will only consider some examples of these different types of arguments.

2.1 The Argument from Different Circumstances

According to this argument, the golden rule does not respect the fact that different people are different and that different individuals might be in different circumstances. There are at least two versions of this argument.

The argument from different external circumstances Suppose that you need to remove your appendix and that you are not a doctor, but that the doctor does not need to remove her appendix. (1) You want the doctor to remove your appendix. (2) If the golden rule is true and you want the doctor to remove your appendix, then you ought to remove the doctor’s appendix. (3) It is not the case that you ought to remove the doctor’s appendix (because the doctor does not need to remove her appendix and you are not a doctor). Hence, (4) the golden rule is not true.¹⁶

The argument from different internal circumstances Suppose that you love to bungee jump, but that your friend does not. (1) You want your friend to buy you a gift card to a bungee jump session as a birthday present. (2) If the golden rule is true and you want your friend to buy you a gift card to a bungee jump session as a birthday present, then you ought to buy your friend a gift card to a bungee jump session as a birthday present. (3) It is not the case that you ought to buy your friend a gift card to a bungee jump session as a birthday present (because she does not want such a present, and you ought to give her something that she would appreciate). Therefore, (4) the golden rule is not true.¹⁷

Cases like these are problematic for LGR and for GGR, but we can use SGGR (or SGR) to avoid them. If the golden rule is interpreted as SGGR, then premise (2) in these arguments cannot be established. Hence, the conclusion does not follow in either case. SGGR does not entail the proposition that if you want the doctor to remove your appendix then you ought to remove the doctor’s appendix, and it does not entail the proposition that if you want your friend to buy you a gift card to a bungee jump session as a birthday present, then you ought to buy your friend a gift card to a bungee jump session as a birthday present.

Suppose SGGR is true. Then if you want it to be the case that if the doctor were in your situation and you were in the doctor’s situation then she would remove your appendix, and you have a good will, then you ought to remove the doctor’s appendix. However, if you are not a doctor and the doctor does not need to remove her appendix, then it is unlikely that you want it to be the case that if the doctor were in your situation (and hence not a doctor) and you were in the doctor’s situation (and hence would not need to remove your appendix) then she would remove your appendix. Again, assume that SGGR is true. Then if you want it to be the case that if your friend were in your situation and you were in your friend’s situation then your friend would buy you a gift card to a bungee jump session as a birthday present, and you have a good will, then you ought to buy your friend a gift card to a bungee jump session as a birthday present. Yet, if you were in your friend’s situation, then you would not want a gift card of this kind. Hence, it is unlikely that you want something like this.

So, neither of the above arguments (nor any similar argument) can be used to refute SGGR (or SGR). The same-situation clause blocks the problematic consequences.

2.2 The Argument from the Agent’s Flawed Desires

If you (the agent) have flawed desires about how you are to be treated, then the golden rule can entail unreasonable conclusions. There are many versions of this argument. I will consider three different kinds.

The argument from unusual or perverted desires The classic example of this kind of argument refers to a masochist and some individual $x$ that is not a masochist. (1) The masochist wants $x$ to inflict pain on him. (2) If the golden rule is true and the masochist wants $x$ to inflict pain on him, then the masochist ought to inflict pain on $x$. (3) It is not the case that the masochist ought to inflict pain on $x$ (because $x$ does not want to have pain inflicted on him). Hence, (4) the golden rule is not true.¹⁸

This argument is a problem for LGR. However, it does not refute SGGR. The same-situation clause blocks the problematic consequence. Premise (2) is not justified if the golden rule is interpreted as SGGR. Suppose that SGGR is true. Then if the masochist wants it to be the case that if $x$ were in the masochist’s situation and the masochist were in $x$’s situation then $x$ would inflict pain on him, and the masochist has a good will, then the masochist ought to inflict pain on $x$. However, since $x$ is not a masochist and does not want to have pain inflicted on him, it is unlikely that the masochist wants it to be the case that if $x$ were in the masochist’s situation and the masochist were in $x$’s situation, then $x$ would inflict pain on him.

The argument from bad, evil or immoral desires Suppose that both $x$ and $y$ are criminals. (1) $y$ wants it to be the case that if $x$ were in $y$’s situation and $y$ were in $x$’s situation then $x$ would help $y$ commit a crime. (2) If the golden rule is true and $y$ wants it to be the case that if $x$ were in $y$’s situation and $y$ were in $x$’s situation then $x$ would help $y$ commit a crime, then $y$ ought to help $x$ to commit a crime. (3) It is not the case that $y$ ought to help $x$ to commit a crime (because neither $x$ nor $y$ ought to commit any crimes). Consequently, (4) the golden rule is not true.¹⁹

We cannot rely on the same-situation clause to block the problematic conclusion in the example above. This is a problem for SGR. Nevertheless, the argument cannot be used to refute SGGR. If the golden rule is interpreted as SGGR, then we can reject premise (2). Accordingly, the conclusion does not follow. We can make the argument valid by replacing (2) by the following premise: If the golden rule is true and $y$ wants it to be the case that if $x$ were in $y$’s situation and $y$ were in $x$’s situation then $x$ would help $y$ commit a crime, and $y$ has a good will, then $y$ ought to help $x$ to commit a crime, and add the following premise: $y$ has a good will. But then we can reject the new premise. Since $y$ is a criminal and wants it to be the case that if $x$ were in $y$’s situation and $y$ were in $x$’s situation then $x$ would help $y$ commit a crime, $y$ does not have a good will. So, we cannot use this argument to show that the golden rule is not true.

The argument from conflicting or inconsistent desires If you have conflicting or inconsistent desires about how someone ought to treat you, we can (sometimes) derive contradictions from the golden rule. Clearly, this is absurd. Here is one example. Suppose you are a doctor and you are wondering whether you should give a patient the bad news that she has cancer. You are very ambivalent about how you want to be treated in a reversed situation. (1) You want it to be the case that if the patient were in your situation and you were in the patient’s situation then the doctor would tell you the truth, and you want it to be the case that if the patient were in your situation and you were in the patient’s situation then the doctor would not tell you the truth. (2) If the golden rule is true and you want it to be the case that if the patient were in your situation and you were in the patient’s situation then the doctor would tell you the truth, then you ought to tell the truth to the patient. (3) If the golden rule is true and you want it to be the case that if the patient were in your situation and you were in the patient’s situation then the doctor would not tell you the truth, then you ought not to tell the truth to the patient. Suppose that (4) the golden rule is true. Then, (5) you ought to tell the truth to the patient and you ought not to tell the truth to the patient. Hence, (6) it ought to be the case that you tell the truth to the patient and that you do not tell the truth to the patient. (7) ‘Ought’ implies ‘can’. (8) If ‘ought’ implies ‘can’, then it ought to be the case that you tell the truth to the patient and that you do not tell the truth to the patient only if it is possible that you tell the truth to the patient and that you do not tell the truth to the patient. (9) It is not possible that you tell the truth to the patient and that you do not tell the truth to the patient. Hence, (10) the golden rule is not true.²⁰

This argument is a problem for SGR. Still, it cannot be used to refute SGGR. If the golden rule is interpreted as SGGR, we can reject steps (2) and (3) in the argument above. Those premises do not include a good will clause. Suppose we replace (2) and (3) with similar premises that do contain a clause of this kind and that we add the premise that you have a good will. Then the new argument is valid. Nevertheless, in that case, we can reject the premise that you have a good will. According to our definition, your will is good only if everything you want ought to be. Therefore, your will is not good if it is inconsistent (given that ‘ought’ implies ‘can’). Consequently, you do not have a good will in this scenario. So, neither this nor any similar argument can be used to refute SGGR.

2.3 The Argument from the Recipient’s Flawed Desires (The Soft Fallacy)

If $x$ (the recipient) has flawed desires, then the golden rule can have undesirable consequences. There are many versions of this kind of argument, since others can have flawed desires in many different ways. I will consider two examples of this kind of problem. The first example is inspired by Immanuel Kant.²¹

The argument from the recipient’s bad desires Suppose that you are a judge and you wonder whether or not you should send this criminal to jail. Furthermore, suppose that the evidence that this criminal is guilty is extremely strong and that he does not want you to send him to jail. (1) If this criminal were in your situation and you were in this criminal’s situation, then you would want it to be the case that the judge does not send you to jail. (2) If the golden rule is true and if it is true that if this criminal were in your situation and you were in this criminal’s situation then you would want it to be the case that the judge does not send you to jail, then you ought not to send this criminal to jail. (3) It is not the case that you ought not to send this criminal to jail (because the evidence against him is extremely strong). Hence, (4) the golden rule is not true.

Here is our second example.

The argument from the recipient’s uninformed desires Suppose you are a parent, and you have a child. Your child does not want you to stop her from playing in this busy street. (1) If your child were in your situation and you were in your child’s situation then you would want it to be the case that your parent does not stop you from playing in this busy street. (2) If the golden rule is true and if it is true that if your child were in your situation and you were in your child’s situation then you would want it to be the case that your parent does not stop you from playing in this busy street, then you ought not to stop your child from playing in this busy street. But this is absurd. (3) It is not the case that you ought not to stop your child from playing in this busy street. In fact, you ought to stop your child (because if you do not, she will be subjected to extreme danger). Hence, (4) the golden rule is not true.²²

These arguments are problematic for PGR. In fact, PGR seems to be equivalent to (or at least entail) the so-called platinum rule: if $x$ wants you to do $H$ to $x$, then you ought to do $H$ to x. If this rule were true, then it would be the case that you should always treat others as they want to be treated. But this is implausible. Still, we can use SGGR to avoid these problems. If the golden rule is interpreted as SGGR, then we can reject premise (2) in both our examples above. According to SGGR, the attitude in the same-situation clause has a wide and not narrow scope. The important question to ask is not what you would want if you were in someone else’s situation, but what you (now) want if you were in someone else’s situation. So, neither of the arguments in this section (nor any similar argument) can be used to refute SGGR.

This concludes our discussion of arguments against the golden rule. The critique of the golden rule is sometimes shallow. The fact that there are many interpretations of this rule that are implausible does not entail that every version of it can be easily refuted.²³

3. The Argument from the Principle of Karma (or Cosmic Justice) for SGGR

In this section, I will show how SGGR can be deduced from a small set of highly plausible premises. I will call the argument in this section ‘the argument from the principle of karma (or cosmic justice) for SGGR’, since it depends on a principle that I will call ‘the principle of karma (or cosmic justice)’.

I will use four premises in this argument. All other steps follow by various standard logical rules. If all these premises are plausible, we have good reason to believe in the conclusion. Here are the premises:

Definition of a ‘good will’ (Premise 1) Individual $y$ has a good will iff everything $y$ wants ought to be the case:$\mathrm{Gy}\leftrightarrow \mathrm{\forall A}\left(\mathrm{WyA}\to \mathrm{OA}\right)$.

We could use the following weaker premise in our argument instead: $y$ has a good will only if everything $y$ wants ought to be. $\mathrm{Gy}\to \mathrm{\forall A}\left(\mathrm{WyA}\to \mathrm{OA}\right)$.

The means-end principle (Premise 2) For every $A$ and $B$, if it ought to be the case that $A$ and it is necessary that if $A$ then $B$, then it ought to be the case that $B$: $\mathrm{\forall A}\mathrm{\forall B}\left(\left(\mathrm{OA}\wedge \text{□}\left(A\to B\right)\right)\to \mathrm{OB}\right)$.

The means-end principle is intuitively plausible, and it can be proved in all so-called alethic-deontic systems where the so-called deontic accessibility relation is included in the so-called alethic accessibility relation. The following semantic argument shows this. According to standard definitions, $\text{□}A$ is true in a possible world $w$ iff $A$ is true in all possible worlds that are alethically accessible from $w$, and $\mathrm{OA}$ is true in a possible world $w$ iff $A$ is true in all possible worlds that are deontically accessible from $w$. Suppose that the means-end principle is not true in a possible world $w$. Then $\mathrm{OA}$ and $\text{□}\left(A\to B\right)$ are true in $w$, while $\mathrm{OB}$ is false in $w$. Since $\mathrm{OB}$ is false in $w$ there is a possible world $w^{\mathrm{\prime }}$ that is deontically accessible from $w$ in which $B$ is false. Hence, $A$ is true in $w^{\mathrm{\prime }}$. Since the deontic accessibility relation is included in the alethic accessibility relation $w^{\mathrm{\prime }}$ is alethically accessible from $w$. Accordingly, $A\to B$ is true in $w^{\mathrm{\prime }}$. Therefore, $B$ is true in$w^{\mathrm{\prime }}$. But this is absurd. It follows that the means-end principle is true. QED

The principle of karma (PK), or the principle of cosmic justice (Premise 3) It is necessary that if you do $H$ to $x$, then if $x$ were in your situation and you were in $x$’s situation, then $x$ would do $H$ to you: $\text{□}\left(\mathrm{Hyx}\to \left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\right)$.²⁴

PK has never before been discussed in the literature as far as I know. So, I would like to comment on it. I call this principle ‘the principle of karma’ or ‘the principle of cosmic justice’ since it is similar to some common ideas about karma or cosmic justice. According to some popular sayings, ‘you reap what you sow’, ‘you get what you give’ and ‘what goes around comes around’. Proverbs of this kind can be found in many cultures and religions all over the world. They are often connected with the idea of some kind of cosmic justice. Exactly how they should be interpreted might differ. When (and where) do we reap (get) what we sow (give)? In this life? In a reincarnated life in the future? In some kind of afterlife? For my purposes, we do not need to answer any questions of this kind.

My principle does not say that we necessarily get what we give in this life or in some other life. Intuitively, it says that you get what you give in an exactly reversed (counterfactual) situation. It says that if you do $H$ to $x$, then in an exactly reversed (counterfactual) situation, $x$ will do $H$ to you. More precisely, it says that it is necessary that if you do $H$ to $x$, then if $x$ were in your situation and you were in $x$’s situation, then $x$ would do $H$ to you. This principle is intuitively reasonable. According to standard definitions, a subjunctive conditional $A\text{□}\to B$ is true in a possible world $w$ iff $B$ is true in all the possible worlds that are as similar to $w$ as possible but in which $A$ is true. So, $\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}$ is true in a possible world $w$ iff $x$ does $H$ to you in all the possible worlds that are as close (or similar) to $w$ as possible but in which $x$ is in your situation and you are in $x$’s situation. Therefore, it is reasonable to assume that if you do $H$ to $x$ in our world, then $x$ will do $H$ to you in every possible world that is as similar to our world as possible but in which $x$ is in your situation and you are in $x$’s situation. Accordingly, it is reasonable to believe that the principle of cosmic justice is true.²⁵

The possibility of role reversals (Premise 4) It is possible that $x$ is in your situation and that you are in $x$’s situation. $\mathrm{♢}\left(S^{y}x\wedge S^{x}y\right)$.

The possibility of role reversals does not entail the proposition that $x$ is in your situation and that you are in $x$’s situation, or that this proposition will be true sometime in the future, or that it is practically, technically, naturally or historically possible. It only says that it is possible. Since a world where you are in your situation and where $x$ is in $x$’s situation is possible, a world where $x$ is in your situation and you are in $x$’s situation also seems to be possible. A possible world where two individuals swap situations is imaginable. Hence, such a world also seems to be possible. It is conceivable that $x$ is in your situation and that you are in $x$’s situation. Consequently, it is reasonable to believe that it is possible that $x$ is in your situation and that you are in $x$’s situation. This does not entail that ordinary human beings can imagine all the details of what it would mean for two individuals to swap places. We do not have to assume that possibility entails conceivability or that conceivability entails possibility.

Note that we do not need the possibility of role reversals as a premise in our argument if we assume the following principle—which appears on line 12 in the deduction of SGGR below—instead of deriving it from PK: It is necessary that if it is true that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, then you do $H$ to x: $\text{□}\left(\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\to \mathrm{Hyx}\right)$. Then we can skip lines 10 and 11 in the proof. This principle is also intuitively reasonable (although I think it might be easier to see that PK is plausible). However, if role reversals are not possible, it is not obvious how interesting our argument is, and we still seem to need the possibility of role reversals, or some similar principle, to be able to defend step 12 in the argument.

We are now in a position to introduce the argument from the principle of karma (or cosmic justice) for SGGR. I will use a natural deduction technique. We begin by assuming the antecedent of the golden rule and derive the consequent. Then I conclude that the whole implication holds, by detaching the assumption. (‘SL’ is an abbreviation of ‘standard (propositional, predicate, modal and counterfactual) logic’; ‘∀E’ stands for ‘universal-elimination’.) Here is the deduction:

1. You want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and you have a good will: $\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\wedge \mathrm{Gy}$. [Assumption.]

2. You want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you: $\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)$. [From 1, by SL.]

3. You have a good will: $\mathrm{Gy}$. [From 1, by SL.]

4. If you have a good will, then everything you want ought to be: $\mathrm{Gy}\to \mathrm{\forall A}\left(\mathrm{WyA}\to \mathrm{OA}\right)$. [By the definition of a good will, Premise 1.]

5. Everything you want ought to be: $\mathrm{\forall A}\left(\mathrm{WyA}\to \mathrm{OA}\right)$. [From 3 and 4, by SL.]

6. If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, then it ought to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you: $\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\to O\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)$. [From 5, by ∀E.]

7. It ought to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you: $O\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)$. [From 2 and 6, by SL.]

8. For every $A$ and $B$, if it ought to be the case that $A$ and it is necessary that if $A$ then $B$, then it ought to be the case that $B$: $\mathrm{\forall A}\mathrm{\forall B}\left(\left(\mathrm{OA}\wedge \text{□}\left(A\to B\right)\right)\to \mathrm{OB}\right)$. [The means-end principle, Premise 2.]

9. If it ought to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and it is necessary that if it is true that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you then you do $H$ to $x$, then it ought to be the case that you do $H$ to x: $\left(O\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\wedge \text{□}\left(\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\to \mathrm{Hyx}\right)\right)\to \mathrm{OHyx}$. [From 8, by ∀E.]

10. It is necessary that if you do not do $H$ to $x$, then if $x$ were in your situation and you were in $x$’s situation then $x$ would not do $H$ to you: $\text{□}\left(\mathrm{\neg Hyx}\to \left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{\neg Hxy}\right)\right)$. [From the principle of karma, PK, Premise 3, and SL.²⁶]

11. It is necessary that if you do not do $H$ to $x$, then it is not the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you: $\text{□}\left(\mathrm{\neg Hyx}\to \mathrm{\neg }\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\right)$. [From 10 and the possibility of role reversals, Premise 4, by SL.]

12. It is necessary that if it is true that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, then you do $H$ to $x$: $\text{□}\left(\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\to \mathrm{Hyx}\right)$. [From 11, by SL.]

13. It ought to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and it is necessary that if it is true that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, then you do $H$ to $x$: $O\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\wedge \text{□}\left(\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\to \mathrm{Hyx}\right)$. [From 7 and 12, by SL.]

14. You ought to do $H$ to $x$: $\mathrm{OHyx}$. [From 9 and 13, by SL.]

15. If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and you have a good will, then you ought to do $H$ to $x$: $\left(\mathrm{Wy}\left(\left(S^{y}x\wedge S^{x}y\right)\text{□}\to \mathrm{Hxy}\right)\wedge \mathrm{Gy}\right)\to \mathrm{OHyx}$. [From 1 to 14, by SL, detaching the assumption.²⁷]

This ends our derivation of the golden rule. The deduction is obviously valid and since the premises are plausible, we have good reason to believe that this version of the golden rule is true. QED

4. Conclusion

In this paper, I have introduced a new version of the golden rule SGGR. According to this version, the principle can be formulated in the following way: If you want it to be the case that if $x$ were in your situation and you were in $x$’s situation then $x$ would do $H$ to you, and you have a good will, then you ought to do $H$ to x. This version is very similar to classical definitions of the rule. I have shown how this principle can be derived from a small set of highly plausible premises and I have defended it against some of the most interesting and/or common arguments against the golden rule that one can find in the literature. I conclude that we have good reason to believe in our rendition of this famous rule.

Acknowledgements

I would like to thank everyone who has commented on this paper for their interesting remarks, questions, and suggestions. This includes two anonymous reviewers for this Journal. I would also like to thank the late Harry J Gensler for a very interesting email exchange on some of the topics in this article.

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

1 For some ‘philosophical’ introductions to this rule, see, for example, Gensler 1996, Wattles 1996, Carson 2010: ch. 6, Carson 2013, and Gensler 2013.

2 See, for example, Hertzler 1934, Wattles 1996, Duxbury 2009, du Roy 2012, and Gensler 2013: ch. 5 for more on the history of the golden rule.

3 For more on the relation between the golden rule and various religions, see, for example, Neusner and Chilton 2008, and Gensler 2013: ch. 3.

4 It has, for example, been accepted (in slightly different forms) by individuals who can be classified as deontologists (Donagan 1977: 57–66), Kantians (Gensler 1996: 12–13, 180), contractualists (Hobbes 1651: ch. XIV, 190), utilitarians (Mill 1863: ch. 2, 28; Sidgwick 1907: 379–80; Hare 1963: 34, ch. 6, 86, 108–9) and natural law theorists (Pufendorf 1672: book 2, 3:13). For some reflections on the relation between the golden rule and the law, see Duxbury 2009.

5 See, for example, Gensler 2013: ch. 2, and Rönnedal 2015. Rönnedal 2015 mentions several questions that are relevant for anyone who wants to define the golden rule precisely. I believe that there might be several different versions of the golden rule that are defensible besides the variant introduced in this paper.

6 Throughout this paper, I will use technical symbols to formalise various important propositions to make it easier to follow the argument. However, it should be possible to understand the various deductions even without any firm grasp of the technical details. I use all symbols in a standard way, for example the truth-functional operators $\mathrm{\neg }$ (negation), ∧ (conjunction), → (material implication) and $\leftrightarrow$ (material equivalence). ‘$\mathrm{\forall }$’ is the universal quantifier that is mainly used for propositional quantification in this paper. ‘$\mathrm{Wy}$’, ‘$O$’, ‘$\text{□}$’ and ‘$\mathrm{♢}$’ are sentential operators. ‘$\mathrm{WyA}$’ is read ‘Individual $y$ wants it to be the case that $A$’, ‘$\mathrm{OA}$’ is read ‘It ought to be the case that $A$’, ‘$\text{□}A$’ is read ‘It is (historically) necessary that $A$’ and ‘$\mathrm{♢}A$’ is read ‘It is possible that $A$’. I assume that ‘$\text{□}$’ and ‘$\mathrm{♢}$’ are so-called (relative) S5-operators. Note that ⬥ is not the dual of $\text{□}$; ⬥ is logically weaker than the dual of $\text{□}$. ‘$\text{□}\to$’ stands for the subjunctive conditional; ‘$A\text{□}\to B$’ is read ‘If $A$ were the case then $B$ would be the case’. The truth-conditions for various sentences are standard. I will assume that the variables in different fundamental formulas are implicitly bound by universal quantifiers and that all fundamental principles are necessary, if not otherwise stated. I will not say anything about how our principles are related to time. Still, we can think of our arguments as implicitly going on in a particular possible world at a particular time. The variable ‘$y$’ is read as ‘you’, and ‘you’ is treated as a variable. For an introduction to modal logic, see, for example, Chellas 1980; for an introduction to deontic logic, see, for example, Gabbay, Parent, et al. 2013, and for more on counterfactual logic, see, for example, Lewis 1973. If $y$ does $H$ to $x$, we shall say that $y$ is ‘the agent’ and that $x$ is ‘the recipient’ (of $H$).

7 ‘SGGR' stands for ‘same-situation, good will golden rule'. I use ‘want’ in an all-things considered, ‘overriding’ sense in this paper. This is compatible with the proposition that some individual may want A (desire A, wish that A) in some weaker sense, even though it is not obligatory that A. So, my use of ‘want' in this paper is partly stipulative. Some may prefer to use the word ‘demand' (or ‘require’) instead of ‘want’ to speak of this kind of attitude. This seems acceptable to me, even though I prefer to use ‘want' in this paper since the golden rule is usually ‘defined' in terms of wants. The ‘ought’ in SGGR should be interpreted as a ‘moral ought’. SGGR is a conditional moral obligation.

8 ‘GGR’ stands for ‘good will golden rule’ and ‘SGR’ stands for ‘same-situation golden rule’.

9 See Gensler 2013: 13, footnote 1. Many other thinkers have defended some clause of this kind; see, for example, Sidgwick 1907: 379–80, Gensler 1996: ch. 5, Carson 2010: ch. 6, and Gensler 2013: 13–15.

10 For more on the wide and narrow scope reading, see, for example, Hare 1963: 108, Hoche 1978, Gensler 2013: 200, and Rönnedal 2015: question 13. PGR is called ‘PGR’ since it is similar to the so-called (literal) platinum rule (‘If x wants you to do $H$ to $x$, then you ought to do $H$ to $x$’).

11 Here is one problem. Among all the properties of an individual $a$ is the property of being identical to $a$. If we include all $a$’s properties in $a$’s situation, then the property of being identical to $a$ is included in $a$’s situation. Then another individual $b$ can be in $a$’s situation only if $b$ can be $a$. But it is not obvious that individuals can swap identities. By only focusing on the universal properties, we avoid potential problems like this.

12 If I try to apply SGGR to decide what to do, should I imagine myself in the other’s external situation (but keep my own likes, dislikes, beliefs, character traits, and so on) or in the other’s exact place (which includes having the other’s likes, dislikes, beliefs, character traits, etc.)? SGGR includes a same-situation clause. According to this, $y$(ou) and $x$ swap all (universal) properties. So, to apply SGGR you should imagine yourself in the other’s exact place (which includes having the other’s likes, dislikes, beliefs, character traits, etc.) as best you can. If we restrict the properties that are swapped to ‘external properties’, SGGR seems to have several counterintuitive consequences (see Gensler 2013: 202 and section 11.2). Note that ‘$S^{y}x$’ does not represent an ordinary relation between $y$ and $x$. So, ‘$S^{y}x$’ should not be replaced by ‘$\mathrm{Syx}$’ in our different formulas. Note also that SGGR is not a decision method or a psychological principle (it is a conditional moral norm). In itself, it doesn’t tell us how we should make various moral decisions. For more on empathy, perspective-taking, putting oneself in someone else’s shoes and similar topics, see, for example, Coplan and Goldie 2011 and Maibom 2017.

13 See, for example, Augustine 400: book 2, ch. 22, Leibniz 1704: book I, ch. II, Gould 1968, Gewirth 1978: 137–38, Gensler 1996: 110–11, and Gensler 2013: 210.

14 When I speak of a ‘good will’ in this paper I mean a ‘morally good will’. Since SGGR is a conditional moral obligation, the ‘ought’ in the definition of a good will should be interpreted as a ‘moral ought’. If you have another view of the meaning of ‘good will’, think of this as a stipulative definition.

15 There are other similar conditions that have similar effects. One could, for example, make a distinction between wants and desires. Everything a person wants is good (and ought to be) according to this reading of the golden rule (see Augustine 400: book 2, ch. 22). One can restrict the wants in the golden rule to wants that are good (whose contents ought to be), etc. (see the references in footnote 13). Some object to a condition of this kind since they think that this makes the golden rule dependent on other moral norms or less useful (see, for example, Bruton 2004). But this argument does not refute our version of the golden rule. SGGR is a conditional norm, not an infallible decision method. One should not expect that it is always easy to use SGGR to decide what one should do. Note that SGGR does not entail that there are no other versions of the golden rule that are true or plausible (and that in some cases may be more ‘useful’). It is, for example, possible to accept SGGR and also accept the golden rule as a rule of thumb (Rönnedal 2015) and as a consistency norm of some kind (Hare 1963; Carson 2010: ch. 6; Gensler 2013).

16 Gensler mentions this example in several places; see, for example, Gensler 1986, 1996: 96, 2002: 234, and 2013: 11.

17 For more information about some similar objections, see, for example, Singer 1963, von Wright 1963: 201, Gewirth 1978, Huang 2005, Gensler 2013: 11–12, 199–200, and Rönnedal 2015.

18 For more on objections of this kind, see, for example Singer 1963, Hoche 1978, Bruton 2004, Carson 2010: 137f, and Gensler 2013: 215–16.

19 See, for example, Sidgwick 1907: 380, Gewirth 1978, and Gensler 2013: 12, which mention similar problems.

20 Problems like this are mentioned by, for example, Whately 1856: 26, Locke 1981, and Gensler 2013: 211–12.

21 See Kant 1785: Ak 4:430 (the footnote on 48). See also Gensler 2013: 209–10.

22 See, for example, Gensler 2013: 16–17 for some similar problems.

23 There are many other objections to the golden rule that I cannot discuss in this paper. For more on various arguments for and/or against this famous principle, see, for example, Kant 1785: Ak 4:430 (48), Cadoux 1912, Hirst 1934, Weiss 1941, Singer 1963, Blackstone 1965, Gewirth 1978, Hoche 1978, Gould 1980, Locke 1981, Gensler 1996: ch. 5, Wattles 1996, Gensler 2002: ch. 11, Bruton 2004, Huang 2005, Carson 2010: ch. 6, du Roy 2012: vol. 2, ch. 13, Gensler 2013: ch. 14, and Rönnedal 2015. However, as far as I am aware there are no objections to the golden rule in the literature that clearly refute SGGR.

24 We interpret the necessity in the means-end principle and in PK as ‘historical necessity’, but our argument for SGGR might perhaps go through on some other readings too (for example, if ‘$\text{□}$’ is read as ‘logical’, ‘metaphysical’ or ‘causal’ necessity). Again, ‘$S^x$’ (‘$S^y$’) is an abbreviation of a conjunction of all properties that $x$ ($y$) in fact has (in a possible world at a particular time) that we are interested in (see Section 1). We focus on the universal properties of $x$ (and $y$). By introducing further restrictions, we obtain different versions of PK. Note that PK does not entail that it is necessary that you do $H$ to $x$ (nor that it is necessary that you do not do $H$ to $x$). PK is compatible with an indeterministic worldview.

25 Note that the second ‘if, then’ in PK is a counterfactual conditional (not a necessary implication). So, this principle does not entail that it is necessary that if you do $H$ to $x$, then it is necessary that x would do $H$ to you in a reversed situation. It is possible that x would not do $H$ to you if x were in your situation, and you were in $x$’s situation even if you do $H$ to x. But this possibility does not refute PK. It is possible that it is true that if A were the case then $B$ would be the case even though it is (logically) possible that A-and-not-B. A counterfactual conditional is weaker than a necessary implication.

26 We assume that ‘$H$’ in PK stands for simple, dyadic action predicates, like ‘help’, ‘tell the truth to’, ‘treat with respect’, etc., and negations of such predicates, for example, ‘not kill’, ‘not lie to’, ‘not rape’, ‘not steal from’, etc.

27 An anonymous reviewer has asked if SGGR is equivalent to the following principle (P): If it ought to be the case that if x were in your situation and you were in $x$’s situation x would do $H$ to you, then you ought to do $H$ to x. The answer to this question is no (unless we assume some other controversial principles). However, it might be interesting to note that (P) follows from the principle of cosmic justice and the possibility of role reversals together with standard logical principles (including the means-end principle). And from (P) and the definition of a good will, we can derive SGGR.