When cognition turns vicious: Heuristics and biases in light of virtue epistemology

Abstract

In this paper, we explore the literature on cognitive heuristics and biases in light of virtue epistemology, specifically highlighting the two major positions—agent-reliabilism and agent-responsibilism (or neo-Aristotelianism)—as they apply to dual systems theories of cognition and the role of motivation in biases. We investigate under which conditions heuristics and biases might be characterized as vicious and conclude that a certain kind of intellectual arrogance can be attributed to an inappropriate reliance on Type 1, or the improper function of Type 2, cognitive processes. By the same token, the proper intervention of Type 2 processes results in the virtuous functioning of our cognitive systems (agent-reliabilism). Moreover, the role of motivation in attenuating cognitive biases and the cultivation of certain epistemic habits (a search for accuracy, being accountable for one's judgments, the use of rules of analysis, and exposure to differing perspectives) points to the tenets of agent-responsibilism in epistemic virtue. We identify the proper use of Type 2 cognitive processes and the habits of mind that attenuate biases as demonstrations of the virtue of intellectual humility. We briefly explore the nature of these habits and the contribution of personality traits, situational pressures, and training in their cultivation.

Keywords::

• Heuristics and Biases
• Intellectual Arrogance
• Intellectual Humility
• Virtue Epistemology

Notes

^[1] Some philosophers worry that psychological theory threatens the viability of virtue epistemology (Alfano, 2012; Olin & Doris, forthcoming). While not aimed at engaging with that literature, this paper does elucidate a surprising harmony between the philosophical and psychological research—a harmony that can, if anything, help dissolve such worries.

^[2] To be sure, one need not be committed to virtue epistemology per se to account for intellectual virtues—someone can account for intellectual virtues without any special loyalty to virtue epistemology—nevertheless, virtue epistemology, naturally enough, offers the most robust and flourishing accounts of intellectual virtue in the philosophical literature.

^[3] As Conee and Feldman have noted, “a fully articulated [process] reliabilist theory must identify with sufficient clarity the nature of the processes it invokes. In doing so, the theory confronts what has come to be known as ‘the generality problem’” (1998, p. 1).

^[4] Also quoted in Pritchard (2005, p. 188).

^[5] See Greco (2003, pp. 356–357), Pritchard (2005, p. 188).

^[6] See chapter 7 of Evans (2007) for a comparative analysis.

^[7] Recently, the language of “systems” has been abandoned by some cognitive scientists in favor of using Type 1 and Type 2 to distinguish these cognitive processes because they do not represent a single system but the many cognitive systems and neural networks that support each type of thinking (Evans & Stanovich, 2013).

^[8] Included in Type 1 processing are implicit learning and conditioning, decision-making rules and principles that are so well practiced as to be automatic, and the regulation of behavior by emotions. Stanovich has given the name TASS (The Autonomous Set of Systems) to these processes because they “respond automatically to triggered stimuli [and are not] under the control of the analytic processing system (System 2)” (2009, p. 57).

^[9] And this means that, according to the credit theory of knowledge espoused by some reliabilist virtue epistemologists (Greco, 2009, 2010, 2012), Type 1 cognition can produce credit-worthy beliefs—beliefs we would deem knowledge—so long as it is employed in a context where it is generally adequate to its task. Also see Axtell (forthcoming).

^[10] As measured by SAT scores and a test of verbal ability, though cognitive ability was also a unique and independent predictor of argument evaluation performance.

^[11] Since much of Type 1 processes are unconscious, we are not accusing the agent of being willfully arrogant. Instead arrogance means to preference the self as a source of information, when what the self knows, by itself, does not provide enough for believing in accordance with the facts, whether the agent is conscious of this preference or not.

^[12] We have defined elsewhere that intellectual humility, as a virtuous mean, can be defined as “holding a belief with the firmness warranted,” which avoids the vice of intellectual arrogance (holding your belief too firmly when it is not warranted) on the one hand, and intellectual diffidence (holding a belief too loosely or giving in to another's belief too soon) on the other (Samuelson, Church, Jarvinen, & Paulus, 2012).

^[13] Insofar as research into heuristics and biases helps us understand intellectual humility, it is useful to think of intellectual humility as the absence or the opposite of intellectual arrogance. As we state above in the introduction, we hold the view that intellectual humility is not simply the opposite, or absence of something (like intellectual arrogance) but that a robust definition should include positive attributes.

^[14] There is some intriguing new research into what is called implicit theory of mind in young children that gives evidence for a type of perspective taking that is akin to Type 1 cognition (i.e., autonomous, fast, and cognitively efficient, though somewhat inflexible; Apperly & Butterfill, 2009). What is less known and more controversial is the relationship between implicit and explicit perspective taking, that is, whether the explicit form is of a different type (akin to Type 2 cognition, Apperly & Butterfill, 2009) or if they both tap into the same cognitive processes in conscious and unconscious ways (Schneider, Lam, Bayliss & Dux, 2012). In either case, perspective taking could be enhanced through experience and practice by making explicit perspective taking so habitual as to become automatic, on the one hand, or by building the capacity of existing implicit forms, on the other, or both.

Additional information

Funding

This work was supported by the John Templeton Foundation [grant number 15628].

Notes on contributors

Peter L. Samuelson

Peter L. Samuelson is a post-doctoral researcher at Fuller Theological Seminary School of Psychology

Ian M. Church

Ian M. Church is a post-doctoral researcher at St. Louis University