Skip to main content
SpringerLink
Log in

Non-Naturalist Moral Realism, Autonomy and Entanglement

  • Published:
Topoi Aims and scope Submit manuscript

Abstract

It was something of a dogma for much of the twentieth century that one cannot validly derive an ought from an is. More generally, it was held that non-normative propositions do not entail normative propositions. Call this thesis about the relation between the natural and the normative Natural-Normative Autonomy (or Autonomy for short). The denial of Autonomy involves the entanglement of the natural with the normative. Naturalism entails entanglement—in fact it entails the most extreme form of entanglement—but entanglement does not entail naturalism. In a ground-breaking paper “The autonomy of ethics” Arthur Prior constructed some intriguing counterexamples to Autonomy. While his counterexamples have convinced few, there is little agreement on what is wrong with them. I present a new analysis of Autonomy, one which is grounded in a general and independently plausible account of subject matters. While Prior’s arguments do establish shallow natural-normative entanglement, this is a consequence of simple logical relationships that hold between just about any two subject matters. It has nothing special to do with the logical structure of normativity or its relation to the natural. Prior’s arguments (along with several others) leave the fundamental idea behind natural-normative Autonomy intact. I offer a new argument for deep entanglement. I show that in any framework adequate for dealing with the natural and the normative spheres, a purely natural proposition entails a purely normative proposition, and vice-versa. But this is no threat to non-naturalist moral realism. In fact it helps ameliorate the excesses of an extreme non-naturalism, delivering a more palatable and plausible position.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. This of course assumes that facts necessitate whatever propositions they make true. This is widely even if not universally endorsed.

  2. This is by no means intended as a watertight argument. Some naturalists have offered more detailed arguments for the systematic violation of Autonomy—e.g. (Jackson 2003, pp. 562–3). I think Jackson’s argument is also flawed. The concrete counterexample to Autonomy that I construct here does not assume naturalism.

  3. Prior (1960).

  4. I am not particularly interested here in entailment relations amongst uninterpreted sentences. We can, of course, entertain the validity of certain schemas, involving variables ranging over interpreted sentences. A schema (such as S∧T ⊨ T∧S) is valid if every substitution instance of the schema is valid.

  5. It may be that a perspicuous language is one in which the structure of a sentence perfectly mirrors the structure of the proposition it expresses, thus reducing the import of the sentence/proposition distinction. See (Tichý 1988) for an extended treatment of this.

  6. Classes can be identified with mappings from a domain to the truth values. So this is tantamount to the thesis that the content of a proposition is a class of world-times. I include partial mappings here, something that is not always embraced by possible-worlds accounts of propositions. If a proposition induces a partial mapping from world-times to truth values, at those world-times at which it is undefined the proposition is truthvalueless.

  7. It will be convenient to also talk of entailment between the contents of propositions: P CONentails Q CON if and only if P entails Q. Since mappings from world-times to truth values are just (possibly partial) classes, this is also tantamount to: P CON⊆ Q CON.

  8. Admittedly it is common for mathematicians to say things like: “the function λy.exp(y + 1,2) contains as constituents both the addition and exponentiation functions”. But here they are not really talking about a function—a mapping from pairs of numbers to numbers. Rather, they are talking about a certain way of getting at, or “constructing”, a function via some other functions. A different way of arriving at the very same function is this: λy.(y×(y + 2) + 1). This latter construction of the function does contain as constituents the addition and multiplication functions, but does not contain exponentiation. See (Tichý 1986) for an informal exposition of the distinction between functions and constructions, and (Tichý 1988) for an extended analysis within transparent intensional logic.

  9. My example is rather gentler than Prior’s. He uses: all New Zealanders ought to be shot.

  10. An entailment schema is valid if every instance of it is valid. Even if the entailment schema is invalid this instance of it might be valid.

  11. See (Mares 2010) for a relevantist construal of Autonomy that renders it compatible with global supervenience, and hence with a version of naturalism.

  12. Mackie (1946).

  13. Mackie (1977).

  14. See (Oddie 2005) Chap. 1.

  15. Negation takes truth values to truth values. As King Lear noted, nothing comes from nothing. One cannot apply a function like negation to nothing at all and expect to end up with a truth value.

  16. Russell’s theory, of course, offers a different solution to the puzzle, and it does entail that the number of hairs on the head of the King of France is neither even nor odd. That this sounds odd is, I think, a defect of the theory but plenty have been able to live with it. However, it does have other more serious defects. Russell’s theory imputes existential import to propositions involving descriptions when the description occurs in a subordinate clause that is not propositional. That I am meditating on the number of hairs on the King of France’s head does not entail that the King of France, or the number of hairs on his head, exists. But on Russell’s account it does. The problem is ineradicable because not all attitudes can be parsed as propositional.

  17. Ought is more plausibly a property of certain properties, but I go here with the flow in deontic logic.

  18. Since ∨ is a truth function and P and Q are not truth values but propositions, the logical form of the disjunction of P and Q is not perspicuously represented as P∨Q or ∨PQ. Where P is a construction of a mapping ƒ from world-times to truth values, let [ P wt ] be the application of ƒ to the pair w,t. If ƒ is defined at w,t then [ P wt ] is a construction of the value of ƒ at w,t. Let [∨P wt Q wt ] be the application of the disjunction function ∨ to whatever pair of truth values (if there is such a pair) constructed by [ P wt ] and [ Q wt ]. Then λw λt [∨P wt Q wt ] gives the logical form of the disjunction of P and Q. λw λt [∨P wt Q wt ] constructs a function g that takes w,t to true whenever both P and Q have truth values at w,t and one of those is true; false whenever both P and Q have truth values at w,t and either of those is false; and is undefined whenever either P or Q fails to yield a truth value at w,t.

  19. Note that if equivalence is taken to be mutual entailment then even with truth value gaps P and True( P ) are equivalent since the following two entailment schemas are valid: P ⊨ True( P ) and True( P ) ⊨ P. Despite this, P and True( P ) may induce distinct mappings from world-times to truth values, since True (P) is false whenever P is truthvalueless. So the following are not valid entailment schemas: ¬True (P) ⊨ ¬P ; ¬True (P) ⊨ True(¬P). In the first case the premise can be true while the conclusion is truthvalueless. In the second case, the premise can be true while the conclusion is false.

  20. If O is truthvalueless so is N∨O. Whenever both premises are true so is the conclusion.

  21. See, for example, the papers in Pigden (2010) by Pigden, Schurz, Mares, and Restall and Russell.

  22. I cannot now locate the source of such examples. Perhaps it was the oral tradition.

  23. Adapted from Nelson (1995).

  24. Pigden 1989 and Schurz 1997. For a summary of their contributions see (Pigden 2010), pp 33–36.

  25. For doubts about this contextualist strategy see (Brown 2014).

  26. This would not be the case if there are truth value gaps. For then O, ¬O and O∨¬O might all fail to have a truth value at some world-time at which N is true. Again, however, we could replace O with True(O) and N with True(N). True(O) is normative, True(N) is natural, and True(O)∨¬True(O) is equivalent to True(N)∨¬True(N).

  27. For an argument against the Boolean closure of the class of natural properties see (Oddie 2005), Chap. 6. If natural properties carve out convex regions of the natural space, then, since negation and disjunction do not preserve convexity, natural properties are not closed under these operations. This is a very strong notion of naturalness—too strong, I think, to capture a reasonable account of Autonomy and Entanglement.

  28. The notion of subject matters as partitions was introduced in Oddie (1986) (108–111) and independently by David Lewis in his (1988), which in turn draws on the earlier work on questions by Belnap and Steele (1976). My analysis is not dissimilar from the one given in Brown (2014). But Brown argues for Autonomy and in the end I will show that it must fail, in any sufficiently rich framework.

  29. Well, almost. There may be no number n such that n numbers the planets in the Solar System —whenever the Solar System does not exist (e.g. one second after the Big Bang). It is not that there are 0 planets in the Solar System at that world-time, but that the question doesn’t even arise at that world-time. It contains a false presupposition. So we need a further element of the partition, one which corresponds to the answer: the question does not arise or does not have a value at the world-time in question.

  30. To preserve the appropriate content for propositions with truth value gaps, for any contents C and D, C⊗D must be undefined at any world-time at which either C or D is undefined.

  31. The distinction between fusions and hybrids was introduced in Oddie and Demetriou (2007).

  32. Pigden makes this objection to a classification similar to the one I am proposing here, although it is not based on subject matters and does not distinguish fusions and hybrids.

  33. For ease of exposition I abbreviate the cumbersome “world-time” to “world” throughout the proof.

  34. See (Oddie 2005, 2016 and 2017).

  35. I am grateful to an anonymous referee for pressing this objection. Obviously it is too far ranging a topic to do full justice to in a single paragraph reply, but I sketch my position on it nevertheless.

  36. For an informal introduction to role theory see (Tichý 1987). For the logic of roles see (Tichý 1988).

  37. This is how (Brown 2014) squeezes Autonomy out of an analysis similar to mine.

  38. I would like to thank the audience members at various presentations of this paper for comments: those at the Prior conference in Christchurch in December 2014; those who attended my talk at the Uppsala Philosophical Society in October 2015; department members at Notre Dame University colloquium in November 2016. I also thank Dom Bailey and two referees at Topoi for many helpful criticisms and corrections.

References

  • Belnap ND, Steel TB (1976) The logic of questions and answers. Princeton University Press, Princeton

    Google Scholar 

  • Brown C (2014) Minding the is-ought gap. J Philos Log 43(1):53–69

    Article  Google Scholar 

  • Jackson F (2003) Cognitivism, a priori deduction, and Moore. Ethics 113:557–575

    Article  Google Scholar 

  • Lewis D (1988) Statements partly about observation. Philos Pap 17(1):1–31

    Article  Google Scholar 

  • Mackie JL (1946) A refutation of morals. Australas J Philos 24:77–90

    Article  Google Scholar 

  • Mackie JL (1977) Ethics: inventing right and wrong. Penguin Books

  • Mares ED (2010) Supervenience and the autonomy of ethics: yet another way in which Relevant Logic is superior to Classical Logic. In Pigden 2010:272–289

  • Nelson MT (1995) Is it always fallacious to derive values from facts?. Argum 9(4):553–562

    Article  Google Scholar 

  • Oddie G (1986) Likeness to truth. Reidel, Dordrecht

    Book  Google Scholar 

  • Oddie G (2005) Value, reality and desire. Oxford University Press, Oxford

    Book  Google Scholar 

  • Oddie G (2016) Fitting attitudes, finkish goods, and value appearances. In: Schafer-Landau Russ (ed) Oxford Studies in Metaethics. Oxford University Press, Oxford, pp 74–101

    Chapter  Google Scholar 

  • Oddie G, Demetriou D (2007) The fictionalist’s attitude problem. Ethical Theor Moral Pract 10(5):485–498

    Article  Google Scholar 

  • Pigden C (1989) Logic and the autonomy of ethics. Australas J Philos 67(2):127–151

    Article  Google Scholar 

  • Pigden C (ed) (2010) Hume on is and ought. Palgrave Macmillan, New York

    Book  Google Scholar 

  • Prior AN (1960) The autonomy of ethics. Aus J Philos 38(3):199–206

    Article  Google Scholar 

  • Restall G, Russell G (2010) Barriers to implication. In Pigden 201:243–259

  • Schafer-Landau (forthcoming). Desire and the good: finding the right fit. In Deonna J, Lauria F The Nature of Desire. Oxford University Press (forthcoming)

  • Schurz G (1997) The is-ought problem: a study in philosophical logic. Kluwer, Dordrecht

    Book  Google Scholar 

  • Tichý P (1986) Constructions. Philos Sci 53(4):514–534

    Article  Google Scholar 

  • Tichý P (1987) Individuals and their roles. In: Svoboda V et al (eds) Pavel Tichý’s collected papers in logic and philosophy. Filosofia and the University of Otago Press 2004, Prague and Dunedin, pp 749–763

  • Tichý P (1988) The foundations of Frege’s logic. De Gruyter, Berlin

    Book  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Philosophy Department, University of Colorado at Boulder, UCB 232, Boulder, CO, 80309-0232, USA

    Graham Oddie

Authors
  1. Graham Oddie
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Graham Oddie.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Oddie, G. Non-Naturalist Moral Realism, Autonomy and Entanglement. Topoi 37, 607–620 (2018). https://doi.org/10.1007/s11245-017-9454-z

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11245-017-9454-z

Keyword

Search

Navigation