Can realists reason with reasons?

ABSTRACT

I argue that realism about reasons is incompatible with the possibility of reasoning with reasons, because realists are committed to the claim that we are aware of reasons by way of ordinary beliefs, whereas a proper understanding of reasoning excludes that our awareness of reasons consists in beliefs. In the first three sections, I set forth five claims that realists standardly make, explain some assumptions I make concerning reasoning, and show why realism, so understood, cannot accommodate the truism that we reason with reasons. I then consider two proposals for how to avoid the problem.

KEYWORDS:

• Reasons
• reasoning
• realism about reasons
• taking condition

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Contemporary philosophers who endorse a version of non-reductive realism include Nagel 1970 and 1986; Scanlon 1998 and 2014; Parfit 1997 and 2011; Wallace 2006; Cuneo 2007 and Enoch 2011.

2 FACT allows for at least two readings, depending on what you understand by a fact. According to a weak version, facts are true propositions. FACT will then follow directly from TRUTH. According to a stronger version, facts are truth-makers of propositions that are prior to and constitutively independent of our representational practices. FACT will then be a substantial further claim. Realists like Parfit or Scanlon seem officially to endorse the weak version, though some of what they say actually seems to involve the strong version. Enoch’s Robust Realism, by contrast, is a version of the stronger claim.

3 I do not claim that TRUTH, FACT and IRREDUCIBILITY are together sufficient for realism about reasons. My only claim is that they are necessarily entailed by it. For instance, error theorists could agree with TRUTH, FACT and IRREDUCIBILITY, but deny an existential claim to the effect that normative facts exist.

4 See, for example, Foot 2000, 53–57; or McDowell 2010, 5–6.

5 Compare, for example, Scanlon 1998, 39; Scanlon 2007, 99; Schroeder 2007, 152–155; Enoch 2011, 225–233. On Enoch’s reading (2011, 226, FN 14), Parfit 1997, 113, also endorses the claim.

6 Those who accept it incluBroome 2013; Boghossian 2014; Valaris 2014; Kietzmann 2018; Müller 2019; Hlobil 2019a.

7 Compare, for example, Gibbons 2009 and Boghossian 2019. McHugh and Way 2016 argue against the connection between the Taking Condition and the agentive character of reasoning. However, Hlobil 2019b offers strong grounds for rejecting their account. Others who argued against the Taking Condition incluKornblith 2012 and Richard 2019. For a defence of the Taking Condition against some of their objections, compare Boghossian 2016, 2018 and 2019.

8 This is not uncommon – compare, for example, Hlobil 2021, 205. Moreover, the assumption seems quite natural. It is commonly assumed that the reason relation is a favouring relation, i.e., if you say that p is a reason for A then you are more or less saying that p favours A. What is its relation to the support relation that we find in reasoning, which the Taking Condition on Boghossian’s formulation is concerned with? In reasoning, support often has to do with logical consequence, such that the truth, say, of the premises guarantees the truth of the conclusion. But notice that support takes this form only in deductive and formally valid inference. Arguably, there are other forms of inference as well, such as abductive reasoning or materially valid inferences. Much scientific reasoning, for example, is not deductive, and much practical reasoning is arguably materially and not formally valid. We should not exclude such kinds of reasoning from our account. In particular, we should not exclude practical reasoning in which intentions are inferred from given circumstances, and scientific reasoning in which beliefs are inferred from evidence. However, the support relations involved in these two kinds of reasoning are nothing other than favouring relations: that circumstances support an intention means that the circumstances favour, or are reasons for, the intention, and that evidence supports a belief means that the evidence favours, or is sufficient reason for, the belief. So at least in the cases I am interested in, cases in which we reason with epistemic or practical reasons, the support relation involved in such reasoning is a reason relation.

9 Although they may not be exact synonyms in ordinary English, most philosophers do not distinguish between ‘to reason’ and ‘to infer’, but use them more or less interchangeably.

10 Carroll’s allusion to Zeno of Elea’s runner paradox suggests that this is how he himself understood the problem. For Zeno’s paradox, see Aristotle 1951, VI.9, 239b14–240a18.

11 For the former suggestion, see Chudnoff 2014, and for the latter, see Kietzmann 2018.

12 See, for example, Broome 2013.

13 See Müller 2019.

14 Compare Mayr 2011, Chapter 5, for an extensive discussion of the problem of deviant causal chains.

15 See Valaris 2014.

16 Thanks to T.M. Scanlon for making this suggestion in correspondence.

17 He shared this idea with me in an email exchange.

18 I presented various ancestors of this paper to audiences in Kloster Rohr, Konstanz, Stuttgart, Utrecht and Zürich. Thanks to the participants for helpful discussion and to Jason Bridges, Magnus Frei, Johann Gudmundsson, Nora Heinzelmann, Ulf Hlobil, Douglas Lavin, Berislav Marušić, Dawa Ometto, T.M. Scanlon, Markos Valaris and two anonymous referees for comments on previous versions. I am also grateful to the Deutsche Forschungsgemeinschaft (DFG) for funding my work on this paper, and to Aaron Shoichet for improving my English.

Additional information

Funding

Work on this paper was supported by Deutsche Forschungsgemeinschaft (DFG projects ‘Aristotelian Constitutivism’ and ‘Capacities and the Good’).

Notes on contributors

Christian Kietzmann

Christian Kietzmann is currently a research fellow at Friedrich-Alexander-Universität Erlangen-Nürnberg. He mainly works on the philosophy of action, meta-ethics and ancient philosophy.