"The importance of awareness for understanding language": Correction to Rabagliati, Robertson, and Carmel (2018)

Abstract

Reports an error in "The importance of awareness for understanding language" by Hugh Rabagliati, Alexander Robertson and David Carmel (Journal of Experimental Psychology: General, 2018[Feb], Vol 147[2], 190-208). In the article, funding details for David Carmel were not included in the author note. The second sentence of the author note should appear as follows: David Carmel was supported by the European Research Council (ERC Advanced Grant XSPECT-DLV-692739, awarded to Andy Clark). The online version of the article has been corrected. (The following abstract of the original article appeared in record 2018-03118-002.) Is consciousness required for high level cognitive processes, or can the unconscious mind perform tasks that are as complex and difficult as, for example, understanding a sentence? Recent work has argued that, yes, the unconscious mind can: Sklar et al. (2012) found that sentences, masked from consciousness using the technique of continuous flash suppression (CFS), broke into awareness more rapidly when their meanings were more unusual or more emotionally negative, even though processing the sentences' meaning required unconsciously combining each word's meaning. This has motivated the important claim that consciousness plays little-to-no functional role in high-level cognitive operations. Here, we aimed to replicate and extend these findings, but instead, across 10 high-powered studies, we found no evidence that the meaning of a phrase or word could be understood without awareness. We did, however, consistently find evidence that low-level perceptual features, such as sentence length and familiarity of alphabet, could be processed unconsciously. Our null findings for sentence processing are corroborated by a meta-analysis that aggregates our studies with the prior literature. We offer a potential explanation for prior positive results through a set of computational simulations, which show how the distributional characteristics of this type of CFS data, in particular its skew and heavy tail, can cause an elevated level of false positive results when common data exclusion criteria are applied. Our findings thus have practical implication for analyzing such data. More importantly, they suggest that consciousness may well be required for high-level cognitive tasks such as understanding language. (PsycINFO Database Record (c) 2019 APA, all rights reserved).