Estimates of nonlinguistic and syntactic factors to account for activations.

For Nested, Simple, Reverse, and Same, the estimates for short and long stimuli were added together, because each factor’s unit load would be invariable between short and long stimuli under each of the sentence and string conditions. Because the matching orders or symbol orders were identical between the Nested and Reverse conditions, the unit load of memory span or counting was invariable between the Nested and Reverse conditions, which was also invariable between the Reverse and Same conditions, thus invariable among the Nested, Simple, Reverse, and Same conditions. For brevity, the contrasts of [Nested – Simple] and [Reverse – Same] are denoted with a double prime mark, i.e., Nested” and Reverse”, respectively. Note that the estimates of memory span in Nested” and Reverse” also became identical, and that the Reverse” contrast makes the listed estimates null, except memory span. The last two syntactic factors, whose models were best in Table 5, consistently accounted for the results of Figure 4F. All estimates of the other factors unlisted here were null in Reverse”, which cannot account for the results of Figure 5C and 5D.