Refining and Testing “Counterintuitiveness” in Virtual Reality: Cross-Cultural Evidence for Recall of Counterintuitive Representations

Abstract

The experiment presented provides partial cross-cultural empirical support for Pascal Boyer's theory of the transmission of minimally counterintuitive (MCI) ideas. Boyer hypothesized that concepts with a small number of counterintuitive features are better remembered and more faithfully communicated than extremely counterintuitive concepts or comparable ordinary or even unusual concepts. This transmission advantage may help to explain the cross-cultural ubiquity of religious/supernatural concepts, which often have counterintuitive features. The experiment was conducted in Second Life, an online 3D virtual world. Fifty English-speaking western participants and 51 Chinese-speaking participants from far-eastern nations viewed intuitive and counterintuitive test items and then were asked to free recall the displays immediately and after varying delays. Results show that MCI displays were not better recalled than intuitive displays at initial reporting. For both samples, however, the amount of time elapsed since exposure to the test items correlated significantly with memory degradation for intuitive concepts but not for counterintuitive concepts. These results suggest that although MCI concepts may not be more easily encoded than intuitive concepts, once they are encoded they may be more easily retrieved than intuitive concepts. Results also show that, among the westerners, increased age predicted poorer delayed recall of MCI but not intuitive items, suggesting that the MCI effect may bear most directly on transmitting ideas to adolescents and young adults.

Notes

¹That is to say, the mind's “crowded zoo” (Cosmides & Tooby, 2000) of innate, or “maturationally natural” (McCauley, 2011), functionally specialized information-processing programs.

²It is important to distinguish counterintuitive concepts from those that violate practiced social behaviours or scripts, or are otherwise bizarre or strange (see Schank, 1999; Schank & Abelson, 1977, for work on enhanced memory for schema violations). Strange events, objects, or properties from an idiosyncratic or cultural perspective are “counterschematic” and not “counterintuitive” (J. L. Barrett, 2008). To illustrate, asking for the bill at a sit-down restaurant before the ordered meal has arrived could be counterschematic but not counterintuitive. Whether something is counterschematic is individually and culturally variable, whereas counterintuitive phenomena should be universal. Previous research (J. L. Barrett & Nyhof, 2001) has shown that mnemonic dynamics of counterintuitive objects are not reducible to distinctiveness effects or merely being “counterschematic.”

³These assumptions may well turn out to be entirely false; MCI test items may generate much more inferential potential than is currently assumed. However, aside from a simple questionnaire proposed by Gregory and Barrett (2009), there currently exists no method by which to measure inferential potential for standardization purposes.

⁴Hypotheses 4 and 5 are not motivated by the central tenets of MCI theory but, rather, by interesting results obtained from previous MCI-related research.

⁵It has been suggested to us that perhaps virtual reality is so different from everyday reality that it does not tap into people's “real-life” cognitions. Throughout the anthropological record, however, the majority of counterintuitive concepts are transmitted through some type of virtual environment (e.g., televisions, newspapers, myths, etc.). Also, even in the most radical forms of virtual reality currently available to researchers—for example, the four-walled cave environments that use headgear, use motion sensors, and allow movement in real space—people react to stimuli (e.g., a hole in the ground; Slater et al., 1995; hostile displays from avatars, Pertaub, Slater, & Barker, 2002) as though they were real.

⁶We have no way to independently verify the age of participants, but this design limitation is most likely to increase the risk of type-II error (failure to find a relationship between age and recall) instead of a type-I error (false positive). If participants lied about their age such that they fell in the wrong age-band and such lying was systemic in either a positive or negative direction, it would at worst compress the age range of participants, making it harder to detect age-effects. If participants lied about their ages randomly, this would introduce noise, making it harder to find an age-effect.

⁷A verbal distraction was used in counterpoint to the visual, iconic nature of the experimental task. It may be that such a task presented greater mnemonic interference than in previous MCI research, but we have no reason to think it would selectively interfere with some items over others.

⁸Because participants were received on a walk-in basis until 50 westerners and 50 far-easterners had successfully completed the project, participants were not counted as participants until completion and we did not keep tally of those sent away for concerns of delay or lag.