Timing Influences on The Specificity of Implied Movement

Timing Influences on The Specificity of Implied Movement Interference Effects Tyler Hubbard, Alan Bisby, Samantha Emerson, John Ibañez, Richard Tillman, & William Langston Middle Tennessee State University Background • This poster is the story of a pursuit; the pursuit of an elusive effect that has remained tantalizingly close, yet forever beyond our grasp. The data from the pursuit are presented below for your consideration. However, we feel that the “news” from this poster is not in the data themselves, but in the problem they pose for theories of embodied cognition. This topic will reappear in the general discussion. • Langston, Hubbard, and Emerson (2009) presented data showing that implied movement interference effects were very specific. Sentences describing motion interacted with moving letters (but not static letters), words describing objects in typical locations interacted with static letters (but not moving letters). An unresolved question related to the nature of these interactions. Basically, matching sentence direction with letter movement produced interference in our data, as in Kaschak et al. (2005). However, matching sentence and movement direction produced facilitation in Glenberg and Kaschak (2002). Borreggine and Kaschak (2006) proposed a timing effect to explain these discrepancies (see also Kaschak et al. for a similar timing account). • Our purpose was to manipulate the onset of letter movement in our motion sentence task to explicitly evaluate this timing account. Initially, we were going to extend this manipulation to the static letters condition because our results contradicted previous research using that task, and we hoped that the timing account could be used to resolve that discrepancy. However, we were unable to move beyond the motion sentences. Experiment 1: Replication with Timing • Eighty undergraduates participated in this experiment. Data from seven participants were removed for low accuracy. • A sentence appeared at the center of the screen. Each critical sentence implied either an upward or downward movement (24 of each; e. g. The balloon rose into the sky, plus 24 filler sentences with no implied movement). (Some of the critical sentences were from Kaschak et al. , 2005; all sentences were normed by a separate group of participants. ) Participants read the sentence for 1500 ms or 3000 ms. Next, an A or L moved up or down. Participants were instructed to push the corresponding letter as soon as they identified it. Participants were then shown a second sentence. They were to judge whether it was the same sentence as the one they saw before. Participants completed ten practice trials before they moved to the actual study. Experiment 2: More Methodological Rigor • Sixty-four undergraduates participated in this experiment. Data from five participants were removed for low accuracy. • The following modifications were made to the materials and procedure: • The sentences were rewritten. Twenty four intransitive verbs describing upward or downward motion were chosen from the Meteyard and Vigliocco (2009) norms, plus 24 verbs that did not have a strong upward or downward implied motion. The verbs were fit into the sentence frame “The <x> <verb>s”. The sentences from the first experiment were a mix of transitive and intransitive verbs, the person varied (e. g. , some second and some third person), and in some of the sentences a prepositional phrase either carried the motion implication or reinforced it. All of these differences affected the data. • On each trial, participants saw the subject of the sentence, pressed the space bar to advance, and then read the verb for a prescribed period of time before the letter moved. The durations were 375 ms, 750 ms, and 1500 ms. This procedure allowed us to better control where in their simulation participants were when the letter moved. 1500 ms presentation 3000 ms presentation 375 ms presentation 750 ms presentation 1500 ms presentation The interaction between sentence movement direction and target movement direction was not significant, F(1, 72) = 1. 46, MSE = 7270, p =. 23. However, the comparison between up and down implied sentence motion for the downward moving targets was significant, t(72) = -2. 34, p =. 02. The comparison between up and down sentence motion for the upward moving targets was not significant, t(72) = -0. 76, p =. 45. The interaction between sentence movement direction and target movement direction was not significant, F(1, 72) = 0. 38, MSE = 3325, p =. 54. The comparison between up and down implied sentence motion for the downward moving targets was significant, t(72) = 2. 66, p =. 01. The comparison between up and down sentence motion for the upward moving targets was not significant, t(72) = 1. 63, p =. 11. The interaction between sentence and target was not significant, F(1, 58) = 0. 94, MSE = 6552, p =. 34. The comparison between up and down sentences for downward moving targets was not significant, t(58) = 1. 66, p =. 10. The comparison between up and down sentences for the upward moving targets was significant, t(58) = 2. 73, p =. 01. The interaction between sentence and target was not significant, F(1, 58) = 0. 38, MSE = 12385, p =. 54. The comparison between up and down sentences for downward moving targets was not significant, t(58) = 0. 71, p =. 48. The comparison between up and down sentences for the upward moving targets was not significant, t(58) = -0. 24, p =. 81. The interaction between sentence and target was marginally significant, F(1, 58) = 3. 68, MSE = 5187, p =. 06. The comparison between up and down sentences for downward moving targets was significant, t(58) = -2. 32, p =. 02. The comparison between up and down sentences for the upward moving targets was not significant, t(58) = 0. 83, p =. 41. 460 500 500 480 480 460 440 460 460 420 440 440 Target Down 400 Target Up 380 Target Down 400 Target Down 420 Target Up 400 400 380 380 360 360 380 360 340 Prime Down Prime Up Figure 1. When the letter moved after 1500 ms there was facilitation if the letter moved down. There was no effect when the letter moved up. 340 Prime Down Prime Up Figure 2. When the letter moved after 3000 ms there was inhibition if the letter moved down. There was no effect when the letter moved up. Experiment 1 Discussion • At first glance, the data appear promising. The significant effect only appears with the downward moving targets, but the same was true for Langston et al. (2009). The facilitation early and inhibition late pattern is consistent with Borreggine and Kaschak’s (2006) result. However, there were two problems: 1) Why no effect for upward moving targets? , 2) Langston et al. found inhibition at 1500 ms letter presentation, not facilitation; why are the results the opposite here? 340 Prime Down Prime Up Figure 3. For 375 ms there was no effect when the letter moved down. There was facilitation when the letter moved up. 340 Prime Down Prime Up Figure 4. For 750 ms there was no effect when the letter moved down and there was no effect when the letter moved up. Prime Down Prime Up Figure 5. For 1500 ms there was facilitation if the letter moved down. There was no effect when the letter moved up. Experiment 2 Discussion • At first glance, the data again appear promising. There is facilitation early; timing is clearly having an effect. However, there are still problems: 1) Why is the facilitation now for upward moving targets? , 2) Why is the effect in the slowest delay facilitation again (instead of inhibition), and why now for downward moving targets? • After a heroic effort to clean up the stimuli from the first experiment and an overhaul of the procedure to guarantee tighter control of the presentation of the stimuli, we are still left with a confusing pattern, and none of the effects have replicated. General Discussion • The obvious question that arises when comparing our data to the extensive literature on embodiment effects in language is: What are we doing wrong? • One possibility is that we aren’t doing anything wrong. Each of the findings, by itself, could be explained with a coherent story, but why isn’t an overall pattern emerging? • Another possibility is that we could attempt even more methodological rigor (adjust the timings, change the procedure some more, etc. ). The problem with this suggestion is that there is no guidance on what the additional methodological rigor should be and why (before seeing the results) those changes should be effective. • Another concern is the lack of a coherent explanation for various clues in the data. For example, the middle timing in Experiment 2 produced the highest variability. Is this meaningful? Why? • We are not proposing that all of the findings reported to date are Type I errors. Rather, we are suggesting that it is time for embodiment research to move from the “any interaction demonstrates that there is an effect” stage to an ability to accurately predict what should happen given a particular set of experimental parameters and to explain problems in the data in a non-post-hoc fashion. • One possibility that we are considering is that individual differences are contaminating the results. Averaging across different kinds of participants is masking overall effects. • We considered reading speed in the studies reported here. However, it did not influence the pattern of results in either study. We are currently evaluating additional individual differences variables. References Borreggine, K. L. , & Kaschak, M. P. (2006). The action-sentence compatibility effect: It’s all in the timing. Cognitive Science, 30, 10971112. Glenberg, A. M. , & Kaschak, M. P. (2002). Grounding language in action. Psychonomic Bulletin and Review, 9, 558 -565. Kaschak, M. P. , Madden, C. J. , Therriault, D. J. , Yaxley, R. H. , Aveyard, M. , Blanchard, A. A. , & Zwaan, R. A. (2005). Perception of motion affects language processing. Cognition, 94, B 79 -B 89. Langston, W. , Hubard, T. , & Emerson, S. (2009, November). The specificity of implied movement interference effects. Poster presented at the 50 th annual meeting of the Psychonomic Society, Boston, MA. Meteyard, L. , & Vigliocco, G. (2009). Verbs in space: Axis and direction of motion norms for 299 English verbs. Behavior Research Methods, 41, 565 -574.