Persuasive as the target article is in arguing that “reasoning proper” is implicitly intended “to devise and evaluate arguments intended to persuade” (abstract of the target article), it says too little about the unconscious “process of inference” that generates the “intuitive beliefs” that are input to this reasoning. This is a serious omission, because one cannot document how reasoning might select and shape arguments without specifying what the inputs to reasoning are. Recent work on spontaneous social inferences (e.g., Uleman et al. Reference Uleman, Saribay and Gonzalez2008) illustrates some of the methods and findings that may fill in this gap.
Spontaneous inferences are unintended, unconscious, practically effortless, typically uncontrollable, and apparently ubiquitous. Most research has been on spontaneous trait inferences (STIs; for an early review, see Uleman et al. Reference Uleman, Newman, Moskowitz and Zanna1996). Consider “John returned the wallet with all the money in it.” When asked to memorize or merely familiarize themselves with such sentences, most people infer that John is honest. This has been shown with cued recall, lexical decisions, probe reaction times, savings-in-relearning, and false-recognition paradigms. People are more likely to assert that “honest” was in the sentence paired with John's photo than the sentence paired with Harry's photo, even though it was not (false recognition; Todorov & Uleman Reference Todorov and Uleman2002). When people are subsequently asked to learn word pairs such as “John–honest,” they do so more readily than “Harry–honest,” even though they no longer recognize which trait-implying sentence described John (savings-in-relearning). And they rate John (in a photo) as more honest, even though they cannot remember what he did or that they made any inference at all (Carlston & Skowronski Reference Carlston and Skowronski2005). So, as Mercier & Sperber (M&S) claim, these unconscious inferences provide the raw material for conscious judgments and presumably for the “reasoning proper” that justifies these judgments.
Spontaneous social inferences are not restricted to traits. There is good evidence that goals and situational (not just trait) causes of behavior are spontaneously inferred. When people read about unjust situations, they spontaneously activate such concepts as “unfair” and “injustice,” but only when they imagine themselves being treated unfairly (Ham & Van den Bos Reference Ham and van den Bos2008). They spontaneously infer causes of largely nonsocial events (Hassin et al. Reference Hassin, Bargh and Uleman2002). In these studies, the texts (or pictures; see Fiedler et al. Reference Fiedler, Schenck, Watling and Menges2005) are pretested by asking people for their conscious inferences. Stimuli that reliably imply whatever is of interest are then tested for spontaneous inferences. The same methods have been used to demonstrate that there are cultural and personality differences in who makes which inferences (see Uleman et al. Reference Uleman, Saribay and Gonzalez2008).
Multiple spontaneous inferences can occur simultaneously to the same stimuli. For example, Ham and Vonk (Reference Ham and Vonk2003) showed that both dispositional and situational inferences during comprehension of a single event (“She got an A on the chemistry exam.” → smart, and → easy). This suggests that, just as Swinney (Reference Swinney1979) found that homonyms (“bank”) initially activate multiple meanings (“money,” “river”) during text comprehension, multiple inferences occur spontaneously during the observation of events, and later selection among them occurs on the basis of wider contexts.
Like many concepts, traits have multiple meanings and uses (Uleman Reference Uleman, Malle and Hodges2005). Traits can function as causal explainations of behavior, or traits can function as simple descriptions of behavior. The same is likely true of other concepts that are activated spontaneously. In explicit dialogue, the pragmatic context in which traits appear allows us to determine their intended meaning and function. But when inferences are spontaneous (i.e., unconscious), no such context exists. Recent research has shown that isolated trait terms function cognitively as causes, not merely as descriptions (Kressel & Uleman Reference Kressel and Uleman2010). And subsequent unpublished work (Kressel Reference Kressel2010) shows that people with stronger implicit (as well as explicit) causal theories of traits' meaning are more likely to make STIs.
Such trait inferences can become associated with the “wrong” actors, in spontaneous trait transference (STT). If person A says something that implies a trait about person B, and only person A is present (or pictured), that trait becomes associated with person A (Skowronski et al. Reference Skowronski, Carlston, Mae and Crawford1998). This does not occur if person B is also pictured, however (Goren & Todorov Reference Goren and Todorov2009). This suggests that spontaneously inferred concepts are easily “bound” to incorrect sources. Thus, events can spontaneously activate a variety of unconscious concepts and associations, all of which provide grist for the “reasoning proper” mill.
Which concepts are activated, and which associations or bindings occur, are context sensitive in other ways. Rim et al. (Reference Rim, Uleman and Trope2009) have shown that, consistent with construal level theory, STI is more likely if the actor is more psychologically distant, either temporally or spatially. People think of distant things more abstractly, and traits are an important kind of social abstraction. Furthermore, unpublished data (Rim et al. Reference Rim, Min, Uleman and Chartrand2010) show that nonconsciously primed goals can shape which inferences occur spontaneously and are bound to actors. Thus, nonconscious goals affect spontaneous inferences in several ways, all outside of awareness.
Finally, research on the logical inferences made during text comprehension goes well beyond bridging and predictive inferences. Lea (Reference Lea1995) showed that deductions according to modus ponens (if p, then q; p; therefore q) occur spontaneously, and Campion (Reference Campion2006) uncovered ways that certain and hypothetical inferences differ. Thus, spontaneous inferences are not limited to the social domain. When stimuli present enough clear information and constraints, both logical and illogical inferences occur (e.g., Rader & Sloutsky Reference Rader and Sloutsky2002).
The formation of “intuitive beliefs” is more complex than the target article describes. Research on spontaneous inferences (social and otherwise) can tell us much about how intuitive beliefs are formed and what they are, before reasoning proper shapes them into persuasive arguments.
Incidentally, the argument that people can distinguish good arguments from bad, based on Petty and Cacioppo's (Reference Petty and Cacioppo1979) persuasion research, is completely circular. They have no principled basis for constructing good versus poor arguments; the arguments are simply pretested to have these properties.
You have
Access
Persuasive as the target article is in arguing that “reasoning proper” is implicitly intended “to devise and evaluate arguments intended to persuade” (abstract of the target article), it says too little about the unconscious “process of inference” that generates the “intuitive beliefs” that are input to this reasoning. This is a serious omission, because one cannot document how reasoning might select and shape arguments without specifying what the inputs to reasoning are. Recent work on spontaneous social inferences (e.g., Uleman et al. Reference Uleman, Saribay and Gonzalez2008) illustrates some of the methods and findings that may fill in this gap.
Spontaneous inferences are unintended, unconscious, practically effortless, typically uncontrollable, and apparently ubiquitous. Most research has been on spontaneous trait inferences (STIs; for an early review, see Uleman et al. Reference Uleman, Newman, Moskowitz and Zanna1996). Consider “John returned the wallet with all the money in it.” When asked to memorize or merely familiarize themselves with such sentences, most people infer that John is honest. This has been shown with cued recall, lexical decisions, probe reaction times, savings-in-relearning, and false-recognition paradigms. People are more likely to assert that “honest” was in the sentence paired with John's photo than the sentence paired with Harry's photo, even though it was not (false recognition; Todorov & Uleman Reference Todorov and Uleman2002). When people are subsequently asked to learn word pairs such as “John–honest,” they do so more readily than “Harry–honest,” even though they no longer recognize which trait-implying sentence described John (savings-in-relearning). And they rate John (in a photo) as more honest, even though they cannot remember what he did or that they made any inference at all (Carlston & Skowronski Reference Carlston and Skowronski2005). So, as Mercier & Sperber (M&S) claim, these unconscious inferences provide the raw material for conscious judgments and presumably for the “reasoning proper” that justifies these judgments.
Spontaneous social inferences are not restricted to traits. There is good evidence that goals and situational (not just trait) causes of behavior are spontaneously inferred. When people read about unjust situations, they spontaneously activate such concepts as “unfair” and “injustice,” but only when they imagine themselves being treated unfairly (Ham & Van den Bos Reference Ham and van den Bos2008). They spontaneously infer causes of largely nonsocial events (Hassin et al. Reference Hassin, Bargh and Uleman2002). In these studies, the texts (or pictures; see Fiedler et al. Reference Fiedler, Schenck, Watling and Menges2005) are pretested by asking people for their conscious inferences. Stimuli that reliably imply whatever is of interest are then tested for spontaneous inferences. The same methods have been used to demonstrate that there are cultural and personality differences in who makes which inferences (see Uleman et al. Reference Uleman, Saribay and Gonzalez2008).
Multiple spontaneous inferences can occur simultaneously to the same stimuli. For example, Ham and Vonk (Reference Ham and Vonk2003) showed that both dispositional and situational inferences during comprehension of a single event (“She got an A on the chemistry exam.” → smart, and → easy). This suggests that, just as Swinney (Reference Swinney1979) found that homonyms (“bank”) initially activate multiple meanings (“money,” “river”) during text comprehension, multiple inferences occur spontaneously during the observation of events, and later selection among them occurs on the basis of wider contexts.
Like many concepts, traits have multiple meanings and uses (Uleman Reference Uleman, Malle and Hodges2005). Traits can function as causal explainations of behavior, or traits can function as simple descriptions of behavior. The same is likely true of other concepts that are activated spontaneously. In explicit dialogue, the pragmatic context in which traits appear allows us to determine their intended meaning and function. But when inferences are spontaneous (i.e., unconscious), no such context exists. Recent research has shown that isolated trait terms function cognitively as causes, not merely as descriptions (Kressel & Uleman Reference Kressel and Uleman2010). And subsequent unpublished work (Kressel Reference Kressel2010) shows that people with stronger implicit (as well as explicit) causal theories of traits' meaning are more likely to make STIs.
Such trait inferences can become associated with the “wrong” actors, in spontaneous trait transference (STT). If person A says something that implies a trait about person B, and only person A is present (or pictured), that trait becomes associated with person A (Skowronski et al. Reference Skowronski, Carlston, Mae and Crawford1998). This does not occur if person B is also pictured, however (Goren & Todorov Reference Goren and Todorov2009). This suggests that spontaneously inferred concepts are easily “bound” to incorrect sources. Thus, events can spontaneously activate a variety of unconscious concepts and associations, all of which provide grist for the “reasoning proper” mill.
Which concepts are activated, and which associations or bindings occur, are context sensitive in other ways. Rim et al. (Reference Rim, Uleman and Trope2009) have shown that, consistent with construal level theory, STI is more likely if the actor is more psychologically distant, either temporally or spatially. People think of distant things more abstractly, and traits are an important kind of social abstraction. Furthermore, unpublished data (Rim et al. Reference Rim, Min, Uleman and Chartrand2010) show that nonconsciously primed goals can shape which inferences occur spontaneously and are bound to actors. Thus, nonconscious goals affect spontaneous inferences in several ways, all outside of awareness.
Finally, research on the logical inferences made during text comprehension goes well beyond bridging and predictive inferences. Lea (Reference Lea1995) showed that deductions according to modus ponens (if p, then q; p; therefore q) occur spontaneously, and Campion (Reference Campion2006) uncovered ways that certain and hypothetical inferences differ. Thus, spontaneous inferences are not limited to the social domain. When stimuli present enough clear information and constraints, both logical and illogical inferences occur (e.g., Rader & Sloutsky Reference Rader and Sloutsky2002).
The formation of “intuitive beliefs” is more complex than the target article describes. Research on spontaneous inferences (social and otherwise) can tell us much about how intuitive beliefs are formed and what they are, before reasoning proper shapes them into persuasive arguments.
Incidentally, the argument that people can distinguish good arguments from bad, based on Petty and Cacioppo's (Reference Petty and Cacioppo1979) persuasion research, is completely circular. They have no principled basis for constructing good versus poor arguments; the arguments are simply pretested to have these properties.