A photographer's self-portrait in a mirror.

Why Critical Reasoning Might Not Require Self-knowledge

I recently reread Tyler Burge’s “Our Entitlement to Self-knowledge” (1996). Burge argues that our capacity for critical reasoning entails a capacity for self-knowledge.

Like a lot of philosophy, this paper is barely connected to the relevant science. So when I find myself disagreeing with the authors’ assumptions, I’m not sure whether the disagreement matters. After all, we might disagree because we have different, unfalsifiable intuitions. But if we disagree about facts, then it matters: one of us is demonstrably wrong. In this post I will articulate my disagreement. I will also try to figure out whether it matters.

One more disclaimer. Frankly, I am not sure what Burge means by ‘critical reasoning’ (more on that in a sec). But Burge treats it as roughly synonymous with ‘reflection’. And that term has been defined sufficiently for our purposes, so I will often use ‘reflection’ and ‘reflective reasoning’ to refer to “critical reasoning” in this post.

1.  Reflection Confers Control and Rationality?

The main thrust of Burge’s paper is that critical reasoning requires self-knowledge. That follows from a modus tollens argument.


p style=”padding-left: 30px;”>P1. If critical reasoning didn’t entail self-knowledge, then critical reasoning would not add rationality to reasoning.
P2: But critical reasoning does add rationality to our reasoning.
C1: So critical reasoning entails self-knowledge.

I am getting that argument from the passage below (p. 101, underlining added), by the way.

If one’s judgments about one’s attitudes or inferences were not reasonable — if one had no epistemic entitlement to themone’s reflection on one’s attitudes and their interrelations could add no rational element to the reasonability of the whole process. But reflection does add a rational element to the reasonability of reasoning. It gives one some rational control over one’s reasoning.

I have two concerns about this argument.

Feeling of Control ⊅ Control and/or Rationality

Suppose that reflection involves conscious access. Conscious access often gives us a feeling of control. And if that feeling is what Burge means by ‘control’, then we are in agreement. But if by ‘control’ Burge means causal (i.e., difference-making) control, then I would want to know why. After all, feeling in control does not entail actually exerting control (Wegner 2007).

But let’s assume that we agree that ‘control’ refers only to a feeling of control. It still wouldn’t follow that feeling in control adds a rational element. After all, we often reason badly even when we feel like we are in control of our reasoning — e.g., political reasoning. So perhaps the feeling of control adds the mere possibility of a rational element. In that case, Burge’s argument would entail only the possibility of self-knowledge.

Domain-general vs. Domain-specific Rationality

Second, I grant that reflection is often useful. But reflection is often (and systematicallynot useful. And that fact is difficult to reject or explain away (see this post for more on that). But even if we did explain it away, ‘useful’ might not entail ‘rational’. So I am not yet inclined to accept the idea that reflection is generally more rational than its unreflective counterparts.

And I am not alone. This is one of the now famous objections to a dogma about reflective reasoning (Stanovich and Evans 2013): Reflective reasoning is rational/useful in some domains and/or for some individuals, but reflection is not generally more rational than unreflective reasoning.

2. Reflective Reasoning is Normally True?

But imagine that my questions about rational control were assuaged. I’d still have reservations about the use of ‘true’ below (102):

If reflective judgments were not normally true, reflection could not add to the rational coherence or add a rational component to the reasonability of the whole process.

What does Burge mean by ‘truth’? Does it matter?

I am not inclined to talk about the epistemic value of reflection in terms of truth. Like Stich in Fragmentation of Reason (1990, chapter 5) I often wonder what people mean by ‘truth’. And, like Stich, I think that if what people mean by ‘truth’ is something like Tarski’s ‘truth’, then I don’t think we actually care as much about that as we do about more pragmatic concerns.

But don’t get me wrong. Pragmatism is consistent with caring about the truth — at least sometimes. For example, in the domains in which ‘truth’ is meaningful and we care most about it, then we should care about the truth-conduciveness of our reasoning. But notice that this kind of concern for truth reduces to pragmatic concerns.

3.  Counterresponses

Burge might respond to me by pointing out that my claims entail that critical reasoning is not possible — and that that is bad. If this was Burge’s response, then two counter-responses would come immediately to mind:

What exactly is ‘critical reasoning’?

After trying to understand ‘critical reasoning’ three times now, I am not sure exactly what it is.

…critical reasoning just is reasoning in which norms of reason apply to how attitudes should be affected partly on the basis of reasoning that derives from judgments about one’s attitudes (102).

…critical reasoning is carried out within a single multi-level point of view (112).

Here’s what seems clear: critical reasoning is a kind of reasoning that involves applying norms to one’s attitudes. But this leaves a lot of relevant questions unanswered.

Does critical reasoning entail being conscious of the norms being applied and the attitudes to which they are applied? It’s not obvious to me that it does. And is knowing these attitudes entailed by (or identical to) merely consciously representing them? Again. It’s not obvious. And (again) does critical reasoning involve actual control over our attitudes or merely a feeling of control? Finally, what the heck is “a single multi-level point of view”?

Depending on Burge’s answers to these questions, I might already believe that critical reasoning is impossible. But until Burge answered these questions, I would be unsure what he thinks is supposed to be impossible on my view.

Is critical reasoning realistic?

Depending on what ‘critical reasoning’ refers to, I might have another question: Why believe that “critical reasoning does occur among us”? (p. 103) After all, if that turns out to be an implausible claim, then I would not be worried about denying it. And part of Burge’s notion of critical reasoning is implausible: that self-knowledge can be rationally immediate. So perhaps I do not need to be worried about the impossibility of critical reasoning.

In critical reasoning, …one’s relation to the known attitudes is rationally immediate: they are part of the perspective of the review itself (113).

Given how the brain works, I do not understand how anything can be immediate.  Every neural process is influenced by contemporaneous processing in nearby (and sometimes not-so-nearby) neurons. So all cognition (and perception) would seem to be inferential and penetrated — i.e., the opposite of immediate. And if reasoning is never immediate, then I’m not sure why it would be rationally immediate.

4. Summary

I appreciate Burge’s explanation of how self-knowledge might be relevant to some sort of reflective reasoning. But I cannot yet agree with the relevance that Burge has in mind. To know whether I agree, I’d need to hear more about what Burge means by ‘control’, ‘truth’, ‘critical reasoning’, and ‘immediate’.

None of these comments are condemnatory, of course. Expecting authors to be perfectly clear is unrealistic. And expecting readers to fully understand authors with just one paper’s worth of writing is also unrealistic. So I am not being disingenuous when I claim that “I need to hear more.”

Related Posts

Published by

Nick Byrd

Nick is a cognitive scientist at Florida State University studying reasoning, wellbeing, and willpower. Check out his blog at byrdnick.com/blog