Abstract
Previous work has suggested that, at least during the Attentional Blink (AB), consciousness of an object is an all-or-none phenomenon (Sergent & Dehaene, 2004). However, the partial awareness hypothesis suggests that this all-or-none property of consciousness should be attributed to consciousness of features (Kouider et al., 2010). The all-or-none hypothesis clearly predicts that when one feature is correctly identified, all other features of the object should also be correctly identified. On the other hand, the partial awareness hypothesis predicts that representations of specific features of objects may reach conscious access independently of each other. We tested these competing predictions in two experiments. In experiment 1, participants (n=16) searched an RSVP for two targets (T1 and T2), and the time between the two targets was manipulated. The T2 task consisted of reporting both identity and color (red, green, or blue). When AB magnitude (ABM; lag 9 - lag 2) was computed using T2 report accuracy conditionalized on correct report of one feature, there was still a significant AB for the other feature (color|identity: ABM=.16, p=.009; identity|color: ABM=.51, p<.001). In experiment 2 (n = 23), the number of potential T2 identity responses was matched to the number of color responses (identity: H,S, or X; color: red, green, or blue). This experiment also included a single target control in which T1 was removed from the stream and participants only reported T2 identity and color. In the dual task condition, when ABM was computed using T2 report accuracy conditionalized on correct report of one feature, an AB was observed for the other feature (color|identity: ABM=.27, p<.001; identity|color: ABM=.22, p<.001). The results of these two experiments support the partial awareness hypothesis, showing that, at least during the AB, there can be a dissociation between consciousness of different features of the same object.
Meeting abstract presented at VSS 2013