Commentary on “From Human Reasoning to Belief”
Lesley Walker
In Joshua Mugg’s book, From Human Reasoning to Belief, he offers an empirically viable alternative to dual process theory and in doing so unifies three contentious cases of belief. The first are cases of wishful thinking, the second, cases of faith, and the third cases often labeled as the result of implicit attitudes. Philosophers have often been reluctant to characterize these cases as paradigmatic beliefs, but Mugg’s Soundboard Account of Human Reasoning explains how each belief can be the result of the same reasoning process despite their seemingly disparate properties.
The soundboard account is fairly intuitive. A soundboard has a variety of knobs, sliders and switches all on the same machine. Switches are binary, while knobs and sliders come in degrees. The position of the soundboard’s knobs and switches dictates the sound of the resultant mix. Mugg suggests we think about reasoning processes and beliefs in a similar manner. Similarly, any reasoning process can operate with a variety of properties. Depending on the external factors of a given task, a reasoning process may be more or less fast, more or less automatic, or involving more or less working memory. Thus, beliefs with various properties arise from reasoning processes with various properties, even though each is produced by a singular system.
One element of Mugg’s view I find most interesting is his notion of control. In his view, while the initial properties of the reasoning process are determined by the external factors of a given task, the subject can rerun and alter the reasoning process by changing modes of operation (e.g., making a fast process slower) or introducing additional cognitive states or heuristics to the process. A reasoning process is more controlled the more times it is rerun, the more modes of operation are changed, and the more new states are brought in.
Mugg suggests that subjects can have a variety of motivations for rerunning a reasoning process. For example, he discusses a case in which his train to Brussels is late and he has a connecting train he doesn’t want to miss. His initial reasoning process outputs the belief that he will not make his connection, but, not wanting to be late, he reruns the process. He might reconsider the speed at which he can walk or recall his confidence (or lack thereof) in the scheduling system of the Amtrak. This altered reasoning process may or may not produce a new belief, either way, it exhibits more control than that initial go.
The possibility of control has interesting implications in the ethics of belief. If belief is entirely involuntary, an ethics of it seems impossible, but the capacity for control makes responsibility possible. Mugg does not argue for any particular set of norms governing belief, but suggests that voluntariness and the involvement of motivation in belief formation may place belief under the practical and moral domains as well as the epistemic domain. I find this compelling, but wonder, to what extent does the possibility of control open us up to an ethics of belief?
Our initial reasoning processes are often set by external factors. Not all of our beliefs exhibit sufficient levels of control—though theoretically we always have the choice of control. I could rerun any reasoning process, and perhaps the fact that I choose not to do so is, in itself, a choice subject to evaluation. This seems intuitive. Many of the reasoning errors we are often concerned with seem to occur due to more automatic processes of belief. Mugg’s case of the self-professed egalitarian, in which said egalitarian sees a White man and a Black man standing beside the valet and hands the Black man his keys, may be one example. In this case, the egalitarian may not be controlling his reasoning process, allowing heuristics and biases to adjudicate belief and behavior. This makes control seem like the solution. If we can control our reasoning processes, we might avoid these sorts of worrying results.
However, interestingly, many of Mugg’s examples of rerunning reasoning processes are examples in which the rerun leads to a less accurate belief—or at the very least in no way guarantees a belief that is better. Mugg explains in Chapter 4 that belief revision is often guided by coherence with one’s existing beliefs rather than truth-conduciveness. Heuristics are prioritized over new information, producing a strong “myside” bias. Likewise, he compares the rerunning of the reasoning process to the rerunning of empirical studies in which scientists aim to achieve significance by either p-hacking, adding more and more subjects to a study until achieving the requisite p-value, or by running the data of a particular study through multiple different statistical models until significance was found. We may rerun reasoning processes as many times as we like, adding or altering whatever is needed, and stopping the inquiry process when our desired result is achieved.
This raises questions about the kind of control Mugg’s account explains. He is explicit that we do not have direct voluntary control over belief itself—only whether and how to rerun the reasoning process. We cannot choose to believe whatever we want, so we only have indirect control over our beliefs. As the examples show, greater control in this sense doesn’t necessarily produce better beliefs. A reasoning process that has been rerun can be just as flawed as the initial go. If that’s right, then the ethical significance of control is less clear. This brings motivation to the forefront. Perhaps we ought to care less about the outcome of a process, and more about one’s reasons for inquiry: whether we’re aiming genuinely at truth, or attempting to reflect moral values, etc. But with motivations doing the normative work, control looks increasingly irrelevant. A person can have good motivations for reasoning regardless of their level of control in the outcome. If that’s the case, then how much does control matter if exercising it doesn’t particularly increase the quality of our beliefs? While this may provide some ground for an ethics of belief, it leaves the state of the debate in a complicated position.
The questions raised here by no means exhaust the intellectual breadth of Mugg’s book. From Human Reasoning to Belief provides a thorough examination of human cognition that captures both the rigidity and flexibility of our reasoning practices. In doing so, he places empirical work on cognition at the heart of some of the most central debates in epistemology and philosophy of mind, offering both a compelling alternative to dual process theory and a cohesive and unique framework with significant implications for the ethics of belief.