One of the more reoccurring sci-fi plot devices is that of traveling to a “twin” earth, where persons are for the most part the same, except that different events have occurred; a recent heated debate in metaethics has been on what occurs when, stipulating that each party has discovered the objective moral truths of the respective worlds, we reach twin earth and a different set of moral facts is objectively true for them. Do our intuitions tell us that when we engage in moral debate with them we are talking past each other, or that we are engaging in a genuine moral debate?
The creators of this thought experiment, Terry Horgan and Mark Timmons (though they are really paying homage to Hilary Putnam’s similar thought experiment), posit that if each group has a different set of moral facts that is true for them then they should be talking past each other, but that our intuitions tell us that we are engaged in a genuine moral debate, such that there are not two sets of moral facts that are true for each group, and this because moral terms do not refer to moral properties, casting doubt on the semantic theories of reference posited by moral realists, and thus the existence of moral properties in general.
The moral realist can deny that we have these intuitions, explain why we have these intuitions in such a way that does not force us to reject the realist’s semantic theory, or accept the intuitions and develop a different semantic theory. Few defenders of naturalistic realism have taken the first option, and most have opted for the second, though some have taken the third approach (notably Geoffrey Sayre-McCord, David Copp, Mark van Roojen and Tristen McPherson). For now I want to contextualize Horgan and Timmons’ argument; I will discuss the responses to their argument in another entry.
Horgan and Timmons’ Argument Against Synthetic Functional Identity Naturalism
1 The Open-Question Argument
G.E. Moore posits that while things can have the property of goodness, the property of goodness, cannot be defined. ‘Good’ cannot be defined in simpler terms because it is, what he calls, a simple notion. Simple notions can be used to explain complex notions, but no simple notion or complex notion can be used to explain a simple notion. Good, as a simple notion, cannot be explained by any other notions, but this does not mean that we do not know what “good” means. Just like the simple notion of yellow, Moore holds that ‘good’ cannot be explained by a more basic notion, but when it is predicated of a subject the listener still understands what ‘good’ or ‘yellow’ means. Thus, simple notions are sufficient unto themselves for explanatory power. ‘Good’, then, is a property that has meaning without being definable by other simple notions, so ‘good’ cannot be identical with or reduced to any other property. As all properties of the natural world can be reduced through scientific theory and the property ‘good’ cannot be reduced, ‘good’ is a non-natural property. Thus, for Moore, to define and reduce the property ‘good’ to another property is to commit the naturalistic fallacy – attempting to define a non-natural property with a natural property.
To support his claim that ‘good’ is not definable, Moore offers what is infamously known as ‘the open question argument’. The argument is powerful in its simplicity. Moore observes that if ‘good’ is identical to another property, then asking whether it is true that ‘good’ is identical to ‘pleasure’ is an odd question to pose because that would be to ask whether a tautology is true. Tautologies must by logical necessity be true, so the question, “is pleasure the good?” must be a closed – that is, not open to debate – question, akin to, “is pleasure is pleasing?” But while the latter question is a closed question, the former seems to be an open question because it is open to debate. If the two properties are identical then the question is a tautology, such that, one can be equally certain that the questions “is pleasure good?” and “is pleasure pleasing?” are true. Moore counters that the certain truth of these claims are far from equal, positing that, “whoever will attentively consider with himself what is actually before his mind when he asks the question ‘Is pleasure (or whatever it may be) after all good?’ can easily satisfy himself that he is not merely wondering whether pleasure is pleasant” (Moore, 418). For Moore, this confirmed that ‘good’ could have meaning while not being reducible to a natural property. However, he did think that a moral property could be co-extensive with a natural property, such that all entities that have the property ‘good’ might have the same natural property.
2 Cornell Realism and the Causal Theory of Reference
With the rise of logical positivism came the decline of cognitivism, and with it moral realism; but in the early 1980’s moral realism returned in the form of the non-reductive naturalism posited by the Cornell realists (among them, Richard Boyd, David Brink, and Nicholas Sturgeon). The Cornell realists posited that moral properties were higher-order functional properties, such that they were not reducible to lower-order properties, but were nevertheless properties of the natural world, like the properties of the social sciences, and were thus natural properties.
The Cornell realists utilized the causal theory of reference posited by Saul Kripke and Hilary Putnam for natural kind terms and applied it to moral terms. Boyd was the principle architect of the semantic theory of Cornell realism, positing that certain mechanisms causally regulate our use of moral terms, whether through actually experiencing moral phenomena, or knowledge of moral terms being socially transmitted and learned. The idea is that because our moral terms are causally regulated by moral properties of actions and states of affairs, our moral knowledge will progress until one day we empirically discover which lower-order properties realize higher-order moral properties such as ‘goodness’.
The parallel to Putnam and Kripke’s theory is that for natural kind terms the essential nature of water as H2o was such that it has the properties of clearness and wetness, and these phenomenal properties regulated our use of the word ‘water’. Even before we knew that water was composed of two hydrogen molecules and one oxygen molecule, when we used the word ‘water’ we unknowingly were also pointing to H2o, because H2o causally regulated our use of the term water.
Thus, if we traveled to a twin earth where everything was the same except that there what twin-earthlings called ‘water’ was composed of XYZ, when we used the word ‘water’ we would not mean what the twin-earthlings meant, because H2o regulates our use of ‘water’, not XYZ, such that, if we engaged in a discussion about ‘water’ with twin-earthlings we would be talking past each other, because ‘water’ refers to different entities for us and them. We would be having a semantic disagreement, not a genuine disagreement about the meaning of ‘water’. Putnam posited the twin-earth water thought experiment to show that the causal theory of reference made a modal necessity claim, such that our terms refer to their corresponding referents in all possible worlds. The Cornell realists posit that because ‘good’ is causally regulated by a set of lower-order properties, when we discover what those properties are we will have discovered a synthetic necessary identity; necessary because the identity relation is true of all possible worlds, but synthetic because it was not analytically obvious in virtue of the meaning of ‘good’ alone.
3 The Strangeness of Moral Supervenience
Horgan and Timmons decry the ease with which Cornell realists have gotten away with positing that moral properties are higher-order properties that irreducibly supervene on non-moral lower order properties by arguing from innocence by association. To Horgan and Timmons, the Cornell realist has gotten away with murder by positing that since it is accepted in the philosophy of mind that the mental supervenes on the physical, that there is nothing mysterious about the Cornell realist holding that the moral supervenes on the non-moral. But this is not a fair comparison, since philosophers of the mind posit supervenience facts that explain how and why the mental supervenes on the physical, while moral realists take supervenience for granted, or so says Horgan and Timmons.
What the Cornell realist owes us, claim Horgan and Timmons, are semantic constraint satisfaction explanations that explain how the higher-order properties exist, and this because semantics constrain and regulate the higher order facts that emerge from the lower order facts. If the moral realist cannot provide semantic constraint satisfaction explanations, then they cannot claim the existence of higher-order moral properties, because if such higher-order moral properties existed then they would causally regulate our moral terms and we would be able to provide semantic explanations for the regulation of their usage.
4 Moral Twin Earth
Horgan and Timmons bade us to imagine, in the distant future, the maiden voyage to moral twin earth, where at that point in time, we have come to conclude that consequentialism is the correct theory of normative ethics, and they have come to the conclusion that deontology is the correct theory of normative ethics. Things go well for a while, but at some point a moral dilemma arises where consequentialism deems an action morally permissible, and deontology deems it impermissible (imagine the organ harvest, the trolley problem, or the judge and the murderous rioters). We have a moral disagreement on our hands, so the question is: what is the nature of this disagreement? Is it merely a verbal or semantic disagreement where the earthlings and the twin-earthlings are talking past each other? Or are the earthlings and the twin-earthlings engaged in a genuine moral disagreement?
If the causal theory of reference is correctly applicable to moral terms, then, like Putnam’s twin earth thought experiment on water, the earthlings and twin earthlings are merely talking past each other when they engage in this moral disagreement. But, say Horgan and Timmons, the moral twin earth thought experiment does not yield this intuition, rather, our intuition is that the earthlings and twin-earthlings are having a genuine and real disagreement about what is morally right.
What this suggests is that there are not enough semantic constraints on moral terms to rigidly fix them for all possible worlds. If there were sufficient semantic constraints then we could explain the difference between ‘good’ for earthlings and ‘good’ for twin-earthlings as a difference in meaning. But our intuition is of a difference in theory and belief, and not meaning, because we do not have the necessary semantic constraints. Thus, the Cornell realist cannot provide the necessary semantic constraint explanations, and therefore, cannot provide a good explanation for causally regulated supervenience relations. Recalling that Horgan and Timmons argued that if moral properties exist then we must be able to provide semantic constraint explanations, and seeing as we cannot provide these semantic constraint explanations, it follows that moral properties do not exist.
5 Conclusion of Part I: Moral Twin Earth as an Open Question Argument
Moore’s open question argument defeated analytic naturalism in the eyes of most philosophers, and Horgan and Timmons intend to show that because a moral twin earth can be created for any synthetic identity put forward by a naturalistic moral realist, Cornell realist or otherwise, it is an open question whether any synthetic necessary identity between a moral term and non-moral properties is true. Horgan and Timmons have thus revived the open question argument in the form of the moral twin earth thought experiment, a variation of the argument from disagreement, and wield it against synthetic naturalists (I am not certain, but am confident that the non-naturalist can escape the dilemma posed by moral twin earth). In a future entry I will explore the naturalists’ responses, beginning with responses that hold that the intuitions generated by moral twin earth are not reliable, and thus, there is no danger to the realist’s use of the causal theory of reference.