Previous studies have validated that participants can distinguish different origins of objects’ shape features, teasing apart features caused by transformation (causal history) from those of the original shape. Considering bite as a transformation example, two experiments were designed to investigate the effect of causal history on the allocation of visual attention. Participants were presented with regular and familiar complete or bitten shapes in Experiment 1 and unfamiliar and irregular complete or bitten shapes in Experiment 2 over a range of stimulus onset asynchronies (SOAs). The task was to identify different probes (i.e., punctuation marks) that equally appeared at four positions around these shapes. The results showed that complete regular shapes had no impact on participants’ reaction times to identify probes that appeared at the four different positions (Experiment 1), whereas complete irregular shapes would facilitate participants’ responses to the probes that appeared at the positions around the “head” of the irregular shape (Experiment 2) regardless of SOAs. When presented with bitten shapes, in the earlier phase of visual processing, participants’ response patterns resembled those found when complete shapes were presented. However, with longer SOAs, participants were faster in identifying probes that appeared at those positions that were around the nontransformed region of the bitten shapes. The results revealed that information about shape features caused by causal history could be incorporated, albeit relatively later, into the allocation of visual attention. The role of causal history in the speculation about one object's future development is discussed.

*causal history*, which refers to the transformations that are applied to an object. The perception of causal history is to infer the process of transformation according to the current shape features. There is evidence showing that observers can distinguish features caused by transformation (causal history) from those that “belong to” the original shape (Schmidt & Fleming, 2018; Schmidt, Phillips, & Fleming, 2019; Schmidt, Spröte, & Fleming, 2016; Spröte & Fleming, 2016).

*f*= 0.25) for the

*F*test on the effect of a four-level within-participant factor (the position, as shown in Design). Twenty-five college students (9 males and 16 females) aged 17–21 years (

*M*= 18.47,

*SD*= 0.92) with normal or corrected-to-normal vision participated in this study for financial compensation. We recruited a few more participants than required in case that some participants’ data could be deleted. They were all right-handed and naive to the purpose of the experiment. The viewing distance was approximately 80 cm.

*SD*from an individual's mean RT (1.2% of trials) were eliminated from further analysis.

*d*was calculated by dividing the mean of the differences between conditions by their standard deviation (Cohen's

*d_z*). The results showed that only the main effect of SOA was significant,

*F*(2.51, 60.18) = 9.85,

*p*< 0.001, η

_{p}

^{2}= 0.291, and Bonferroni-corrected pairwise comparisons revealed that participants were faster to identify the probe when the SOA was 75 ms (

*M*= 570.52 ms,

*SD*= 148.23) compared to 255 ms (

*M*= 526.10 ms,

*SD*= 129.54,

*t*(24) = 4.12,

*p*= 0.004,

*d_z*= 0.82), 450 ms (

*M*= 516.95 ms,

*SD*= 108.97,

*t*(24) = 4.23,

*p*= 0.003,

*d_z*= 0.84), and 750 ms (

*M*= 497.54 ms,

*SD*= 92.89,

*t*(24) = 4.48,

*p*= 0.002,

*d_z*= 0.90), except for 135 ms (

*M*= 544.53 ms,

*SD*= 140.65). The results indicated that complete circle or square shapes do not have a significant influence on the allocation of visual attention.

*F*(1.50, 36.10) = 12.45,

*p*< 0.001, η

_{p}

^{2}= 0.342, and post hoc comparisons with Bonferroni's correction revealed a longer RT mean with the historical position (

*M*= 585.47 ms,

*SD*= 180.41) than with the opposite-history position (

*M*= 548.47 ms,

*SD*= 143.49,

*t*(24) = 4.12,

*p*= 0.002,

*d_z*= 0.82), the orthogonal P1 (

*M*= 540.47 ms,

*SD*=130.42,

*t*(24) = 3.62,

*p*= 0.008,

*d_z*= 0.72), and orthogonal P2 (

*M*= 538.06 ms,

*SD*= 133.63,

*t*(24) = 4.10,

*p*= 0.002,

*d_z*= 0.82). The main effect of SOA was also significant,

*F*(1.32, 31.64) = 15.43,

*p*< 0.001, η

_{p}

^{2}= 0.391, and post hoc comparisons with Bonferroni's correction revealed a longer RT with 75 ms (

*M*= 589.23 ms,

*SD*= 177.80) than 255 ms (

*M*= 547.43 ms,

*SD*= 149.55,

*t*(24) = 5.62,

*p*< 0.001,

*d_z*= 1.12), 450 ms (

*M*= 532.93 ms,

*SD*= 127.19,

*t*(24) = 4.47,

*p*= 0.002,

*d_z*= 0.89), and 750 ms (

*M*= 519.52 ms,

*SD*= 110.93,

*t*(24) = 4.22,

*p*= 0.003,

*d_z*= 0.84), except for 135 ms (

*M*= 576.47 ms,

*SD*= 171.57). In addition, longer RTs of 135 ms than 255 ms (

*t*(24) = 4.56,

*p*= 0.001,

*d_z*= 0.91), 450 ms (

*t*(24) = 4.02,

*p*= 0.005,

*d_z*= 0.80), and 750 ms (

*t*(24) = 3.69,

*p*= 0.012,

*d_z*= 0.74) were also found. These effects were qualified by a significant interaction,

*F*(3.73, 89.44) = 5.41,

*p*= 0.001, η

_{p}

^{2}= 0.184. Simple effect analysis showed that there was no significant effect of position when the SOAs were 75 (

*F*(3, 22) = 2.29,

*p*> 0.10), 135 (

*F*(3, 22) = 1.36,

*p*> 0.28), and 255 ms (

*F*(3, 22) = 2.64,

*p*> 0.07). When the SOA was 450 ms, the effect of position was significant,

*F*(3, 22) = 4.18,

*p*= 0.018, η

_{p}

^{2}= 0.363, and the historical position (

*M*= 596.25 ms,

*SD*= 199.97) resulted in a longer RT mean than the opposite-history position (

*M*= 513.89 ms,

*SD*= 119.47,

*t*(24) = 3.57,

*p*= 0.009,

*d_z*= 0.71), orthogonal P1 (

*M*= 515.86 ms,

*SD*= 108.38,

*t*(24) = 3.56,

*p*= 0.009,

*d_z*= 0.71), and orthogonal P2 (

*M*= 505.73 ms,

*SD*= 104.20,

*t*(24) = 3.64,

*p*= 0.008,

*d_z*= 0.73). When the SOA was 750 ms, the effect of position was also significant,

*F*(3, 22) = 5.40,

*p*= 0.006, η

_{p}

^{2}= 0.424. The historical position (

*M*= 582.11 ms,

*SD*= 184.85) resulted in a longer RT than the opposite-history position (

*M*= 486.70 ms,

*SD*= 91.61,

*t*(24) = 4.00,

*p*= 0.003,

*d_z*= 0.80), orthogonal P1 (

*M*= 505.13 ms,

*SD*= 102.99,

*t*(24) = 3.03,

*p*= 0.035,

*d_z*= 0.61), and orthogonal P2 (

*M*= 504.13 ms,

*SD*= 97.61,

*t*(24) = 3.19,

*p*= 0.002,

*d_z*= 0.64).

*M*= 24.04,

*SD*= 2.03) with normal or corrected-to-normal vision participated in this study for financial compensation. They were all right-handed and naive to the purpose of this experiment. The sample size was calculated as in Experiment 1a.

*SD*from an individual's mean reaction time (1.16% of trials) were eliminated from further analysis.

*F*(1.57, 39.35) = 10.72,

*p*= 0.001, η

_{p}

^{2}= 0.300, and post hoc comparisons with Bonferroni's correction revealed that 75 ms (

*M*= 557.37 ms ,

*SD*= 67.26) resulted in a longer RT mean than 135 ms (

*M*= 527.25 ms,

*SD*= 64.76,

*t*(25) = 6.39,

*p*< 0.001,

*d_z*= 1.25), 255 ms (

*M*= 506.43 ms,

*SD*= 57.40,

*t*(25) = 6.20,

*p*< 0.001,

*d_z*= 1.22), 450 ms (

*M*= 502.80 ms,

*SD*= 68.26,

*t*(25) = 4.24,

*p*= 0.003,

*d_z*= 0.83), and 750 ms (

*M*= 492.01 ms,

*SD*= 75.18,

*t*(25) = 4.19,

*p*= 0.003,

*d_z*= 0.82).

*F*(1.91, 47.63) = 10.03,

*p*< 0.001, η

_{p}

^{2}= 0.286, and post hoc comparisons with Bonferroni's correction revealed a longer RT mean with the historical position (

*M*= 559.97 ms,

*SD*= 82.43) than with the opposite-history position (

*M*= 533.14 ms,

*SD*= 68.04,

*t*(25) = 3.14,

*p*= 0.025,

*d_z*= 0.62), the orthogonal P1 (

*M*= 515.84 ms,

*SD*= 50.78,

*t*(25) = 3.97,

*p*= 0.003,

*d_z*= 0.78), and orthogonal P2 (

*M*= 517.64 ms,

*SD*= 59.39,

*t*(25) = 3.61,

*p*= 0.008,

*d_z*= 0.71). The main effect of the SOA was also significant,

*F*(1.37, 34.12) = 15.52,

*p*< 0.001, η

_{p}

^{2}= 0.383. Post hoc comparisons with Bonferroni's correction revealed that 75 ms (

*M*= 571.53 ms,

*SD*= 79.05) resulted in a longer RT than 255 ms (

*M*= 534.93 ms,

*SD*= 68.15,

*t*(25) = 5.32,

*p*< 0.001,

*d_z*= 1.04), 450 ms (

*M*= 499.32 ms,

*SD*= 62.40,

*t*(25) = 4.41,

*p*= 0.002,

*d_z*= 0.87), and 750 ms (

*M*= 494.64 ms,

*SD*= 68.88,

*t*(25) = 4.35,

*p*= 0.002,

*d_z*= 0.85), except for 135 ms (

*M*= 557.82 ms,

*SD*= 78.78). In addition, 135 ms resulted in a longer RT mean than 255 ms (

*t*(25) = 3.34,

*p*= 0.026,

*d_z*= 0.65), 450 ms (

*t*(25) = 4.04,

*p*= 0.004,

*d_z*= 0.79), and 750 ms (

*t*(25) = 3.90,

*p*= 0.006,

*d_z*= 0.76), and 255 ms resulted in a longer RT mean than 450 ms (

*t*(25) = 3.12,

*p*= 0.044,

*d_z*= 0.61) and 750 ms (

*t*(25) = 3.13,

*p*= 0.046,

*d_z*= 0.61). These effects were qualified by a significant interaction,

*F*(4.76, 118.91) = 4.00,

*p*< 0.001, η

_{p}

^{2}= 0.138. Simple effect analysis showed that there was no significant difference among these positions when the SOAs were 75 ms (

*F*(3, 23) = 3.20,

*p*= 0.042, η

_{p}

^{2}= 0.294

^{1}) and 135 ms (

*F*(3, 23) = 1.23,

*p*> 0.32). The effect of position was significant when the SOA was 255 ms,

*F*(3, 23) = 3.84,

*p*= 0.023, η

_{p}

^{2}= 0.334, and the historical position (

*M*= 565.13 ms,

*SD*= 88.93) resulted in a longer RT than the orthogonal P1 (

*M*= 521.11 ms,

*SD*= 73.46,

*t*(25) = 3.12,

*p*= 0.027,

*d_z*= 0.61) and orthogonal P2 (

*M*= 515.70 ms,

*SD*= 72.20,

*t*(25) = 3.53,

*p*= 0.015,

*d_z*= 0.66). When the SOA was 450 ms, the effect of position was also significant,

*F*(3, 23) = 4.79,

*p*= 0.010, η

_{p}

^{2}= 0.384, and the participants were slowest in recognizing the marks that appeared at the historical position (

*M*= 542.32 ms,

*SD*= 87.19) than at the opposite-history position (

*M*= 484.65 ms,

*SD*= 81.34,

*t*(25) = 3.45,

*p*= 0.012,

*d_z*= 0.68), orthogonal P1 (

*M*= 479.21 ms,

*SD*= 62.78,

*t*(25) = 3.56,

*p*= 0.009,

*d_z*= 0.70), and orthogonal P2 (

*M*= 491.11 ms,

*SD*=70.70,

*t*(25) = 2.87,

*p*= 0.049,

*d_z*= 0.56). The effect of position was significant when SOA was 750 ms,

*F*(3, 23) = 7.36,

*p*= 0.001, η

_{p}

^{2}= 0.490. The participants were slowest in recognizing the marks that appeared at the historical position (

*M*= 547.64,

*SD*= 83.26) than at the opposite-history position (

*M*= 489.56 ms,

*SD*= 93.07,

*t*(25) = 2.96,

*p*= 0.040,

*d_z*= 0.58), orthogonal P1 (

*M*= 461.29 ms,

*SD*= 73.38,

*t*(25) = 4.44,

*p*= 0.001,

*d_z*= 0.87), and orthogonal P2 (

*M*= 480.05 ms,

*SD*= 90.13,

*t*(25) = 3.17,

*p*= 0.024,

*d_z*= 0.62).

*M*= 21.48,

*SD*= 2.29) with normal or corrected-to-normal vision participated in this study for financial compensation. They were all right-handed and naive to the purpose of this experiment. The sample size was calculated as in Experiment 1a.

*3*SD from an individual's mean reaction time (1.06% of total trials) were eliminated from further analysis. Categorization task results showed that the causal origins of shapes were perceived as we intended: Complete shapes were undoubtedly perceived as complete (

*M*= 7.00,

*SD*= 0), bitten shapes were also perceived as bitten (

*M*= 6.00,

*SD*= 1.56), and imposed shapes were perceived as imposed (

*M*= 4.84,

*SD*= 1.75). The rating data were submitted to a one-way ANOVA with shape type on three levels (complete, bitten, impose), and the results revealed a significant main effect of shape type,

*F*(2, 74) = 18.57,

*p*< 0.001. Multiple comparisons with Bonferroni's correction showed that the ratings of complete shapes were higher than that of bitten and imposed shapes (

*p*s < 0.001), but the rating of bitten and imposed shapes did not differ between each other. Therefore, the participants interpreted similar jagged outlines according to different causal origins.

*F*(3, 96) = 0.34,

*p*= 0.80.

*F*(1, 24) = 2.18,

*p*> 0.15. The main effect of position was significant,

*F*(1.35, 32.56) = 9.90,

*p*= 0.002, η

_{p}

^{2}= 0.292, qualified by a significant interaction effect of shape and position,

*F*(1.51, 36.22) = 11.42,

*p*< 0.001, η

_{p}

^{2}= 0.322. Simple effect analysis with Bonferroni's correction showed that when the participants were presented with bitten shapes, they were slower to identify the probe that appeared at the historical position (

*M*= 516.34 ms,

*SD*= 125.99) than that at the opposite-history position (

*M*= 452.25 ms,

*SD*= 57.62,

*t*(24) = 3.27,

*p*= 0.019,

*d_z*= 0.65), orthogonal P1 (

*M*= 425.89 ms,

*SD*= 50.12,

*t*(24) = 3.92,

*p*= 0.004,

*d_z*= 0.78), and orthogonal P2 (

*M*= 432.85 ms,

*SD*= 53.61,

*t*(24) = 3.65,

*p*= 0.008,

*d_z*= 0.73). In addition, the probe that appeared at orthogonal P1 was identified significantly faster than that at the opposite-history position,

*t*(24) = 3.24,

*p*= 0.021,

*d_z*= 0.65. When the participants were presented with imposed shapes, they were significantly slower to identify the probe that appeared at the opposite-history position (

*M*= 508.13 ms,

*SD*= 111.27) than that at the historical position (

*M*= 459.61 ms,

*SD*= 74.05,

*t*(24) = 3.67,

*p*= 0.007,

*d_z*= 0.73) and orthogonal P1 (

*M*= 433.69 ms,

*SD*= 55.77,

*t*(24) = 3.83,

*p*= 0.005,

*d_z*= 0.77) and were slightly slower than that at orthogonal P2 (

*M*= 448.99 ms,

*SD*= 76.33,

*p*= 0.074).

*M*= 19,

*SD*= 1.26) with normal or corrected-to-normal vision participated in this study for financial compensation. They were all right-handed and naive to the purpose of this experiment. The sample size was calculated as in Experiment 1a.

*SD*from an individual's mean reaction time (1.36% of trials) were eliminated from further analysis.

*F*(1.19, 29.84) = 10.82,

*p*< 0.001, η

_{p}

^{2}= 0.302, and post hoc comparisons with Bonferroni's correction revealed that the consistent position (

*M*= 611.11 ms,

*SD*= 134.94) resulted in a longer RT mean than the inconsistent position (

*M*= 523.91 ms,

*SD*= 74.24,

*t*(25) = 3.75,

*p*= 0.006,

*d_z*= 0.73), orthogonal P1 (

*M*= 538.56 ms,

*SD*= 97.93,

*t*(25) = 3.20,

*p =*0.022,

*d_z*= 0.63), and orthogonal P2 (

*M*= 544.26 ms,

*SD*= 104.07,

*t*(25) = 3.00,

*p*= 0.036,

*d_z*= 0.59). The main effect of the SOA was also significant,

*F*(4, 100) = 13.79,

*p*< 0.001, η

_{p}

^{2}= 0.356, and post hoc comparisons with Bonferroni's correction revealed that 75 ms (

*M*= 578.50 ms,

*SD*= 90.89) resulted in a longer RT mean than 135 ms (

*M*= 556.41 ms,

*SD*= 97.50,

*t*(25) = 4.40,

*p*= 0.002,

*d_z*= 0.86), 255 ms (

*M*= 551.07 ms,

*SD*= 101.59,

*t*(25) = 4.40,

*p*= 0.002,

*d_z*= 0.86), 450 ms (

*M*= 541.62 ms,

*SD*= 83.48,

*t*(25) = 6.80,

*p*< 0.001,

*d_z*= 1.33), and 750 ms (

*M*= 544.71 ms,

*SD*= 90.60,

*t*(25) = 5.58,

*p*< 0.001,

*d_z*= 1.09). These effects were qualified by a slightly significant interaction,

*F*(12, 300) = 2.12,

*p*= 0.015, η

_{p}

^{2}= 0.078. Simple analysis showed that, when the SOAs were 75 ms and 450 ms, the difference between the consistent and orthogonal P2 did not reach significance, which was different from what was revealed by the main effect of position.

*F*(1.59, 39.84) = 8.62,

*p*= 0.002, η

_{p}

^{2}= 0.256, and post hoc comparisons with Bonferroni's correction showed that the opposite-history position (

*M*= 539.84 ms,

*SD*= 80.27) resulted in a shorter RT than the consistent position (

*M*= 591.36 ms,

*SD*= 116.56,

*t*(25) = 3.62,

*p*= 0.008,

*d_z*= 0.71) and historical position (

*M*= 585.28 ms,

*SD*= 113.94,

*t*(25) = 3.60,

*p*= 0.008,

*d_z*= 0.71). The main effect of the SOA was also significant,

*F*(1.56, 39.06) = 14.70,

*p*< 0.001, η

_{p}

^{2}= 0.370, and post hoc comparisons with Bonferroni's correction showed that 75 ms (

*M*= 603.23 ms,

*SD*= 132.38) resulted in a longer RT than 135 ms (

*M*= 579.46 ms,

*SD*= 113.92,

*t*(25) = 4.10,

*p*= 0.004,

*d_z*= 0.80), 255 ms (

*M*= 569.65 ms,

*SD*= 102.02,

*t*(25) = 3.86,

*p*= 0.007,

*d_z*= 0.76), 450 ms (

*M*= 548.41 ms,

*SD*= 81.86,

*t*(25) = 4.51,

*p*= 0.001,

*d_z*= 0.88), and 750 ms (

*M*= 540.58 ms,

*SD*= 72.42,

*t*(25) = 4.27,

*p*= 0.002,

*d_z*= 0.84). In addition, 135 ms resulted in a longer RT than 450 ms (

*t*(25) = 3.56,

*p*= 0.015,

*d_z*= 0.70) and 750 ms (

*t*(25) = 3.47,

*p*= 0.019,

*d_z*= 0.68), and 255 ms resulted in a longer RT than 450 ms (

*t*(25) = 3.25,

*p*= 0.033,

*d_z*= 0.64) and 750 ms (

*t*(25) = 3.71,

*p*= 0.010,

*d_z*= 0.73). These effects were qualified by a significant interaction,

*F*(3.10, 77.55) = 4.35,

*p*= 0.006, η

_{p}

^{2}= 0.148. Simple analysis revealed that there was no significant effect of position on RT when SOAs were 75 ms (

*F*(3, 23) = 1.90,

*p*> 0.15) and 135 ms (

*F*(3, 23) = 2.51,

*p*> 0.08). Position had a significant effect on RT when the SOA was 255 ms,

*F*(3, 23) = 4.90,

*p*= 0.009, η

_{p}

^{2}= 0.390, and the opposite-history position (

*M*= 539.45 ms,

*SD*= 92.91) resulted in a shorter RT than the consistent position (

*M*= 612.36 ms,

*SD*= 143.53,

*t*(25) = 3.21,

*p*= 0.022,

*d_z*= 0.63). When the SOA was 450 ms, the effect of position on RT was significant,

*F*(3, 23) = 4.42,

*p*= 0.014, η

_{p}

^{2}= 0.366, and the opposite-history position (

*M*= 513.27 ms,

*SD*= 53.05) resulted in a shorter RT than the consistent position (

*M*= 557.07 ms,

*SD*= 97.07,

*t*(25) = 3.43,

*p*= 0.013,

*d_z*= 0.67) and historical position (

*M*= 577.69 ms,

*SD*= 120.57,

*t*(25) = 3.06,

*p*= 0.031,

*d_z*= 0.60). The effect of position was also significant when the SOA was 750 ms,

*F*(3, 23) = 5.63,

*p*= 0.005, η

_{p}

^{2}= 0.423. The historical position (

*M*= 605.57 ms,

*SD*= 126.33) resulted in a longer RT than the consistent position (

*M*= 533.01 ms,

*SD*= 99.05,

*t*(25) = 3.04,

*p*= 0.033,

*d_z*= 0.60), the inconsistent position (

*M*= 521.47 ms,

*SD*= 87.61,

*t*(25) = 3.68,

*p*= 0.007,

*d_z*= 0.72), and the opposite-history position (

*M*= 502.28 ms,

*SD*= 52.93,

*t*(25) = 4.21,

*p*= 0.002,

*d_z*= 0.83).

*M*= 23.64,

*SD*= 2.34) with normal or corrected-to-normal vision participated in this study for financial compensation. They were all right-handed and naive to the purpose of this experiment. The sample size was calculated as in Experiment 1a.

*SD*from an individual's mean reaction time (1.21% of trials) were eliminated from further analysis.

*F*(1.46, 35.03) = 17.39,

*p*< 0.001, η

_{p}

^{2}= 0.420, and post hoc comparisons with Bonferroni's correction revealed that the inconsistent position (

*M*= 599.62 ms,

*SD*= 111.38) resulted in a longer RT than the consistent position (

*M*= 516.99 ms,

*SD*= 46.04,

*t*(24) = 4.70,

*p*= 0.001,

*d_z*= 0.94), orthogonal P1 (

*M*= 513.46 ms,

*SD*= 51.45,

*t*(24) = 4.55,

*p*= 0.001,

*d_z*= 0.91), and orthogonal P2 (

*M*= 517.57 ms,

*SD*= 54.59,

*t*(24) = 4.35,

*p*= 0.001,

*d_z*= 0.87). The main effect of the SOA was also significant,

*F*(1.53, 36.59) = 37.98,

*p*< 0.001, η

_{p}

^{2}= 0.613, and post hoc comparisons with Bonferroni's correction revealed that 75 ms (

*M*= 585.88 ms,

*SD*= 78.73) resulted in a longer RT than 135 ms (

*M*= 547.95 ms,

*SD*= 68.18,

*t*(24) = 6.09,

*p*< 0.001,

*d_z*= 1.22), 255 ms (

*M*= 528.47 ms,

*SD*= 59.66,

*t*(24) = 8.27,

*p*< 0.001,

*d_z*= 1.65), 450 ms (

*M*= 510.51 ms,

*SD*= 47.57,

*t*(24) = 7.45,

*p*< 0.001,

*d_z*= 1.49), and 750 ms (

*M*= 498.15 ms,

*SD*= 41.25,

*t*(24) = 6.98,

*p*< 0.001,

*d_z*= 1.40). In addition, 135 ms resulted in a longer RT than 255 ms (

*t*(24) = 4.23,

*p*< 0.001,

*d_z*= 0.85), 450 ms (

*t*(24) = 5.11,

*p*< 0.001,

*d_z*= 1.02), and 750 ms (

*t*(24) = 4.97,

*p*< 0.001,

*d_z*= 0.99), and 255 ms also resulted in a longer RT than 750 ms (

*t*(24) = 3.89,

*p*= 0.007,

*d_z*= 0.78). The interaction of position and SOA did not reach significance,

*F*(6.77, 162.65) = 1.49,

*p*> 0.17.

*F*(2.07, 49.62) = 15.34,

*p*< 0.001, η

_{p}

^{2}= 0.390, and post hoc comparisons with Bonferroni's correction revealed that the consistent position (

*M*= 537.13 ms,

*SD*= 55.66) resulted in a shorter RT mean than the inconsistent position (

*M*= 571.83 ms,

*SD*= 88.88,

*t*(24) = 3.42,

*p*= 0.013,

*d_z*= 0.68) and historical position (

*M*= 568.69 ms,

*SD*= 76.75,

*t*(24) = 3.40,

*p*= 0.014,

*d_z*= 0.68). The opposite-history position (

*M*= 514.79 ms,

*SD*= 45.79) resulted in a shorter RT mean than the inconsistent position (

*t*(24) = 4.36,

*p*= 0.001,

*d_z*= 0.86) and historical position (

*t*(24) = 5.54,

*p*< 0.001,

*d_z*= 1.11). The main effect of SOA was also significant,

*F*(1.16, 27.81) = 35.24,

*p*< 0.001, η

_{p}

^{2}= 0.595, and post hoc comparisons with Bonferroni's correction revealed that 75 ms (

*M*= 610.10 ms,

*SD*= 102.84) resulted in a longer RT mean than 135 ms (

*M*= 592.05 ms,

*SD*= 93.91,

*t*(24) = 5.09,

*p*< 0.001,

*d_z*= 1.02), 255 ms (

*M*= 569.15 ms,

*SD*= 80.24,

*t*(24) = 5.96,

*p*< 0.001,

*d_z*= 1.19), 450 ms (

*M*= 490.96 ms,

*SD*= 45.05,

*t*(24) = 6.41,

*p*< 0.001,

*d_z*= 1.28), and 750 ms (

*M*= 478.28 ms,

*SD*= 39.22,

*t*(24) = 6.39,

*p*< 0.001,

*d_z*= 1.28). Additionally, 135 ms resulted in a longer RT mean than 255 ms (

*t*(24) = 4.13,

*p*= 0.004,

*d_z*= 0.83), 450 ms (

*t*(24) = 5.84,

*p*< 0.001,

*d_z*= 1.17), and 750 ms (

*t*(24) = 5.94,

*p*< 0.001,

*d_z*= 1.19), and 255 ms resulted in a longer RT mean than 450 ms (

*t*(24) = 5.25,

*p*= 0.004,

*d_z*= 1.05) and 750 ms (

*t*(24) = 5.58,

*p*< 0.001,

*d_z*= 1.12). These effects were qualified by a significant interaction,

*F*(3.34, 80.25) = 15.08,

*p*< 0.001, η

_{p}

^{2}= 0.386. Simple analysis revealed that the effects of position were significant when SOAs were 75 ms (

*F*(3, 22) = 5.65,

*p*= 0.005, η

_{p}

^{2}= 0.435), 135 ms (

*F*(3, 22) = 6.84,

*p*= 0.002, η

_{p}

^{2}= 0.482), and 255 ms (

*F*(3, 22) = 5.59,

*p*= 0.005, η

_{p}

^{2}= 0.433). The inconsistent position resulted in a longer RT mean than any of the others (

*p*< 0.01 in each case) under the three SOA conditions. The position also had a significant effect on RT when the SOA was 450 ms,

*F*(3, 22) = 7.70,

*p*= 0.001, η

_{p}

^{2}= 0.513. The difference in RT between the opposite-history position (

*M*= 443.98 ms,

*SD*= 65.25) and the consistent position (

*M*= 481.55 ms,

*SD*= 52.07) approached significance,

*p*= 0.057. In addition, the historical position (

*M*= 561.08 ms,

*SD*= 83.67) resulted in a longer RT mean than the consistent position (

*M*= 481.55,

*SD*= 52.07,

*t*(24) = 4.80,

*p*< 0.001,

*d_z*= 0.96), inconsistent position (

*M*= 477.21 ms,

*SD*= 69.45,

*t*(24) = 4.20,

*p*= 0.002,

*d_z*= 0.84), and opposite-history position (

*M*= 443.98 ms,

*SD*= 65.25,

*t*(24) = 4.83,

*p*< 0.001,

*d_z*= 0.97). When the SOA was 750 ms, the effect of position was significant,

*F*(3, 22) = 13.44

*p*< 0.001, η

_{p}

^{2}= 0.647. The opposite-history position resulted in a significantly shorter RT (

*M*= 435.70 ms,

*SD*= 53.44) than the consistent position (

*M*= 471.58 ms,

*SD*= 39.80,

*t*(24) = 4.12,

*p*= 0.002,

*d_z*= 0.82). In addition, the historical position (

*M*= 558.80 ms,

*SD*= 73.24) resulted in a significantly longer RT mean than the consistent position (

*M*= 471.58 ms,

*SD*= 39.80,

*t*(24) = 5.53,

*p*< 0.001,

*d_z*= 1.11), inconsistent position (

*M*= 447.03 ms,

*SD*= 68.16,

*t*(24) = 6.17,

*p*< 0.001,

*d_z*= 1.23), and opposite-history position (

*M*= 435.70 ms,

*SD*= 53.44,

*t*(24) = 6.26,

*p*< 0.001,

*d_z*= 1.25).

*p*= 0.060.

*Nature,*434, 79–83, doi:10.1038/nature03271.

*Experimental Brain Research,*180, 85–96, doi:10.1007/s00221-006-0840-9.

*Psychological Science,*27, 923–930, doi:10.1177/0956797616628525.

*Psychonomic Bulletin & Review,*26, 1266–1272, doi:10.3758/s13423-019-01607-8.

*The Quarterly Journal of Experimental Psychology,*61, 860–868, doi:10.1080/17470210701623738.

*Towards a theory of thinking*(pp. 73–101). Berlin, Germany: Springer.

*British Journal for the Philosophy of Science,*68, v42, doi:10.1093/bjps/axv042.

*Journal of Experimental Psychology Human Perception & Performance,*29, 1211–1227, doi:10.1037/0096-1523.29.6.1211.

*Behavior Research Methods,*39, 175–191, doi:10.3758/BF03193146.

*Journal of Vision,*19(4), 15, doi:10.1167/19.4.15.

*Journal of Experimental Psychology: Human Perception & Performance,*30, 319–329, doi:10.1037/0096-1523.30.2.319.

*Ecological Psychology,*7, 217–256, doi:10.1207/s15326969eco0703_3.

*Spatial Cognition & Computation,*1, 1–29, doi:10.1023/a:1010073227203.

*Journal of Memory and Language,*38, 1–27, doi:10.1006/jmla.1997.2533.

*Cognitive Science,*13, 357–387, doi:10.1016/0364-0213(89)90017-7.

*Journal of Vision,*17(1), 20, doi:10.1167/17.1.20.

*Proceedings of the National Academy of Sciences,*117, 117335–11743, doi:10.1073/pnas.1917565117.

*Nature Neuroence,*6, 82–89, doi:10.1038/nn984.

*Cognitive Neuropsychology,*23, 583–605, doi:10.1080/02643290500310962.

*Quarterly Journal of Experimental Psychology,*64, 669–688, doi:10.1080/17470218.2010.520086.

*Journal of Experimental Psychology: General,*121, 95–111, doi:10.1037/0096-3445.121.1.95.

*Journal of Vision,*18(1), 14, https://doi.org/10.1167/18.1.14.

*Art & Perception,*8, 407–433, doi:10.1163/22134913-20191116.

*Cognitive Psychology,*90, 48–70, doi:10.1016/j.cogpsych.2016.08.002.

*PLoS One,*13, e202115, doi:10.1371/journal.pone.0202115.

*Journal of Vision,*20(6), 2, https://doi.org/10.1167/jov.20.6.2.

*Journal of Vision,*17(3), 18, doi:10.1167/17.3.18.

*Cognition,*189, 167–180, doi:10.1016/j.cognition.2019.04.006.

*Vision Research,*126, 318–329, doi:10.1016/j.visres.2015.04.011.

*Journal of Experimental Psychology—General,*143, 434–454, doi:10.1037/a0032353.

*Journal of Vision,*13(14), 3, doi:10.1167/13.14.3.

*Vision Research,*126, 330–346, doi:10.1016/j.visres.2015.08.009.

*Scientific Reports,*6, 36245, doi:10.1038/srep36245.

*Psychonomic Bulletin & Review,*9, 314–318, doi:10.3758/BF03196287.

*Perception & Psychophysics,*70, 77–87, doi:10.3758/PP.70.1.77.

*Journal of Neuroscience,*38, 8526–8537, doi:10.1523/JNEUROSCI.2800-17.2018.