Inferential self-knowledge reimagined

ABSTRACT

In the epistemology of self-knowledge, Inferentialism is the view that one’s current mental states are normally known to one through inferences from evidence. This view is often taken to conflict with widespread claims about normally-acquired self-knowledge, namely that it is privileged (essentially more secure than knowledge of others’ minds) and peculiar (obtained in a way that fundamentally differs from how others know your mind). In this paper I argue that Inferentialism can be reconceived so as to no longer conflict with these claims. On a positive framing, Inferentialism earns a seat at the table alongside other theories of self-knowledge that take its potentially privileged and peculiar status seriously. On a less positive framing, Inferentialism’s plausibility now hinges on the plausibility of viewing self-knowledge as at least peculiar, and perhaps as privileged as well.

KEYWORDS:

• Self-knowledge
• inference
• Evidential Groundlessness
• privileged access
• peculiar access

Acknowledgements

Thanks are due to Lukas Schwengerer, Lisa Doerksen, and Adam Andreotta for reading and commenting on earlier versions of this article. Thanks are also due to audience members at the Canadian Philosophical Association 2023 Annual Meeting, where Lisa also presented insightful commentary. Finally, I extend special thanks to my anonymous reviewers for feedback that greatly improved this project.

Disclosure statement

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

Notes

1. Another influential view appeals to inferences that are not based on evidence (Byrne, 2018). This is not an Inferentialist view in my intended sense, and hence is beyond my scope here (though see Winokur, 2021a for some discussion).

2. In fact, Inferentialists say the same of self-knowledge regarding one’s past and future mental states. I will focus on self-knowledge of current mental states, just as Inferentialists typically do. This is because Inferentialism as a view about self-knowledge of one’s current mental states is more controversial; more theorists about self-knowledge agree that knowledge of one’s past or future mental states is inferential. Note, furthermore, that by “current” I do not necessarily mean “occurrent” (where “occurrent” contrasts with “standing”).

3. See Byrne (2005, 2018), whose formulations of Privilege and Peculiarity are like my own in several respects. Some will find that this formulation of Peculiarity begs the question against “constitutivist” accounts that take the relevant body of self-knowledge to have no “way” in which it is acquired. If one likes, one can reconstrue Peculiarity so that it reads more inclusively, e.g: in normal cases, others know one’s mind fundamentally differently than one knows one’s mind. Some may also wish to hold to a stronger thesis, according to which self-knowledge is differently grounded than all other forms of empirical knowledge, not just knowledge of other minds. I do not explore this stronger thesis here.

4. Cassam defends the normative claim as follows.

“If I ask you where you think Napoleon was born and you tell me he was born in Corsica then, unless I have some reason to doubt your sincerity, I thereby come to know where you think Napoleon was born. In cases like this it might be tempting to say that I hear your belief in your words and that my knowledge of your belief is in this sense ‘direct’. This temptation should be resisted. Talk of ‘hearing’ your belief is at best metaphorical, and I know you believe that Napoleon was born in Corsica only if I’m justified in believing on the basis of your utterance that this is what you believe. My justification is inferential and so, therefore, is my knowledge.” (2017: p. 735)

5. Even the exceptionally “safe” self-knowledge-conferring inferences described by Byrne (2018) may go wrong – see Fricke (2018, p. 49) and Winokur (2021a).

6. Though see Coliva (2016, p. 87) and Roche (2018, p. 646) for criticisms of this view.

7. Pluralism about normal self-knowledge is an increasingly popular position (e.g., Boyle, 2009; Coliva, 2016).

8. Some philosophers make stronger claims. For example, Donald Davidson writes that “the self-attributer does not normally base his claims on evidence or observation, nor does it normally make sense to ask the self-attributer why he believes he has the beliefs, desires or intentions he claims to have” (1984: p. 103). I work with a weaker, appearances-based thesis about self-knowledge’s evidential groundlessness because it does not presuppose the falsity of Inferentialism, and because it is harder for Inferentialists to deny these appearance claims than their factive variants.

9. If Moran (2004) is right, we are entitled to think that our deliberations settle our attitudes. But this entitlement merely betrays the fact that the actual determination is not itself transparent to us. What we are entitled to is a presumption that such determination obtains.

10. I owe this and the following Inferentialist-friendly reply to an anonymous reviewer.

11. Thanks to an anonymous reviewer for the objection. See Doyle (2019) and Winokur (2023) for more about making one’s self-ascriptions intelligible to oneself.

12. There is a risk, at play in Boghossian’s discussion, of conflating the personal/subpersonal distinction with the system 1/system 2 cognition distinction (Kahneman, 2011; Stanovich & West, 2000). I take no stance on the potential differences between them. Thanks to an anonymous reviewer for forcing this clarification.

13. More recently, it has been proposed that the “taking” might be contemporaneous with the inference itself, rather than serving as a causal antecedent to the inference (Blake-Turner, 2022). I find this proposal congenial, though it does not affect the substance of my discussion. The taking phenomenon itself is what concerns me, not its precise temporal location in the inferential process.

14. Though see Hunter (ms) for discussion of the possibility of voluntarily though non-intentional inference.

15. Setiya (2013) advocates for skepticism about cognitive agency, though I follow Parrott (2015) in holding that the very moderate conception of cognitive agency outlined here is sufficiently uncontroversial.

16. As aforementioned, one might think that we have perceptual knowledge of other minds – maybe we can simply see our friend’s anger, or hear it in their voice (alternatively: you might see that or hear that somebody is in a mental state – cf. Dretske, 1973). Suppose that seeing X and hearing X are factive. In such a case, these perceptual achievements do not provide mere evidence for the truth of what one believes about another person’s mind. But even if this is so, self-knowledge and other-knowledge will remain contrasted with respect to the fact that the latter but not the former is generally expected to receive some sort of articulable epistemic support, whether factive or merely evidential. The difference is precisely what we would expect if one species were normally grounded in subpersonal inferences lacking agential involvement and the other were normally grounded in either agentially-involved person-level inferences or straightforward perception.

17. A somewhat tangential objection to my arguments takes advantage of an idea drawn from Natalia Waights Hickman (2021) in a different argumentative context. They countenance a species of person-level knowledge that is not articulable, such as our knowledge of many of the complicated rules for combinations of word-meanings: “[e]xplicit formulations of such knowledge … [cannot] be recognized as apt or otherwise by the subject, and cannot be cited in explanations or justifications of their actions and attitudes except by third parties.” (Ibid.: p. 711). This, they say, makes it implicit knowledge but, crucially, not subpersonal. The knowledge is still person-level because it rationalizes her intentional linguistic activity. Perhaps, then, inferences normally involved in producing self-knowledge are also person-level processes, albeit non-articulable. And now the objection is that, since they are person-level, they do not differ in their constitutive form from the inferences normally involved in generating knowledge of other minds.

But why suppose that the inferences involved in self-knowledge are person-level yet too complex to articulate? This is our first issue. We have some antecedent sense that knowledge of the vast sea of semantic rules undergirding our speech competencies are too complex to be articulated, whereas it is not obvious that self-knowledge is normally based on inferences in the first place. Moreover, even if we grant that there is such a thing as person-level inference that is not articulable and person-level inference that is, such a difference may still lead to normative asymmetries between self-knowledge and other-knowledge. For instance, while self-knowledge and other-knowledge may both be inferentially justified, the latter seems to be uniquely subject to defeat in the event that the agent fails to articulate the grounds for her inference.

18. For some discussion of the controversy, see Malmgren (2019) and Winokur (2021b, §3.5).

19. Cassam points out that “[b]eing too reflective and critical can slow you down and lead to poorer decision-making than fast or unconscious thinking” (2014: p. 216). Perhaps it is also true that reflective and critical thinking can lead to poorer theoretical commitments than fast or unconscious thinking.

20. Thanks to an anonymous reviewer for distinguishing between person-level confabulatory inferences and person-level self-ascriptions that were caused by sub-personal inferences.

Additional information

Funding

The author has no funding to declare.