Verb gapping: an action-gap compatibility study.

Acta Psychologica 156 (2015) 104–113

Contents lists available at ScienceDirect

Acta Psychologica journal homepage: www.elsevier.com/ locate/actpsy

Verb gapping: An action-gap compatibility study Berry Claus ⁎ Department of German Studies and Linguistics, Humboldt-Universität zu Berlin, Unter den Linden 6, 10099 Berlin, Germany

a r t i c l e

i n f o

Available online 4 August 2014 PsycINFO classification: 2340 cognitive processes 2720 linguistics & language & speech Keywords: Language comprehension Verb gapping Action–sentence compatibility Action simulation Embodiment

a b s t r a c t This study addresses the processing of verb-gapping sentences, e.g., John closes a juice bottle and Jim [ ] a lemonade bottle. The goal was to explore if there would be an interaction between language comprehension and motor action not only for overt action verbs but also for gapped verbs. Participants read gapping sentences that either described clockwise or counter-clockwise manual rotations (e.g., closes vs. opens a juice bottle). Adopting a paradigm developed by Zwaan and Taylor (2006), sentence presentation was frame-by-frame. Participants proceeded from frame to frame by turning a knob either clockwise or counter clockwise. Analyses of the frame reading-times yielded a significant effect of compatibility between the linguistically conveyed action and the knob turning for the overt-verb (e.g., closes/opens a juice bottle) as well as for the gapped-verb frame (e.g., a lemonade bottle) – with longer reading times in the match condition than in the mismatch condition – but not for any of the other frames (e.g., and Jim). The results are promising in providing novel evidence for the real-time reactivation of gapped verbs and in suggesting that action simulation is not bound to the processing of overt verbs. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Can language be “a surrogate for experience” (Taylor & Tversky, 1992, p. 495)? According to some theories of language comprehension (e.g., Barsalou, 1999; Glenberg, 1997, 2007; Zwaan, 2004), the answer to this question should be absolutely positive. These theories assume that language comprehension involves embodied mental simulations that are grounded in perception and action. They share the assumption that representations derived from linguistic input recruit the same modality-specific mental subsystems as representations derived from direct experience. In what follows, I will use the term embodiedsimulation view to embrace different theoretical approaches that all assume a common representational system for linguistic cognition, non-linguistic cognition, and perception and action. The embodied-simulation view is empirically supported by numerous impressive findings (for overviews see e.g., Barsalou, 2008; Meteyard, Cuadrado, Bahrami, & Vigliocco, 2012). Neuroscientific studies have revealed a considerable overlap between the pattern of brain activation that occurs when a particular linguistic expression is processed and the pattern of activation that is involved in directly perceiving the object or doing the activity, which the linguistic expression denotes (e.g., González et al., 2006; Pulvermüller & Hauk, 2006; Tettamanti et al., 2005). Findings from behavioral studies point to an ⁎ Humboldt-Universität zu Berlin, Institut für deutsche Sprache und Linguistik, Unter den Linden 6, 10099 Berlin, Germany. Tel.: +49 30 2093 9673. E-mail address: [email protected].

http://dx.doi.org/10.1016/j.actpsy.2014.07.009 0001-6918/© 2014 Elsevier B.V. All rights reserved.

equivalence between representations derived from linguistic input and representations being used in perception and action by indicating that effects that have been demonstrated in studies on perception and action also show up in language comprehension tasks (e.g., Chen & Bargh, 1999; Claus & Kelter, 2009; Glenberg, Sato, & Cattaneo, 2008; Pecher, Zeelenberg, & Barsalou, 2003). Furthermore, there are several behavioral studies that revealed interaction effects between language comprehension and concurrent perceptual processes (e.g., Bergen, Lindsay, Matlock, & Narayanan, 2007; Kaschak et al., 2005; Meteyard, Bahrami, & Vigliocco, 2007; Zwaan, Madden, Yaxley, & Aveyard, 2004). One important line of evidence – which also is of immediate relevance to the study to be presented in the present paper – stems from studies addressing the interaction between language comprehension and action and indicating that processing linguistic input can affect motor actions. In particular, many studies revealed an action–sentence compatibility effect (Glenberg & Kaschak, 2002), i.e., response times for sentences describing an action in a particular direction were found to be affected by whether the direction of the hand movement required for responding was compatible or incompatible to the direction implied by the sentence (e.g., Bergen & Wheeler, 2005; de Vega, Moreno, & Castillo, 2013; Glenberg & Kaschak, 2002; Zwaan & Taylor, 2006; see also Bub & Masson, 2010). By now, the embodied-simulation view has gained considerable importance in research on language comprehension. It has drawn attention to many new questions leading to novel findings, particularly as to the relationship between language processing and perceptual and motor processes. Yet, the embodied-simulation view is still in its infancy

B. Claus / Acta Psychologica 156 (2015) 104–113

stage. There are several fundamental issues that have largely been neglected so far, in particular genuine linguistic phenomena such as scope, anaphora, and elliptical constructions. The goal of the present study was to explore whether the embodiedsimulation view can contribute to research on the processing of a particular type of elliptical constructions, namely verb gapping (Ross, 1970). Verb gapping is the omission of the finite verb (and possibly additional verbal material) in the second conjunct of a conjoined sentence. An example is given in (1), where the finite verb ordered is elided in the second conjunct. 1. Spencer ordered a cup of coffee and Katherine a cup of tea. Processing a verb-gapping sentence such as (1) requires detecting the gap and retrieving the omitted verb information, with the missing information being given in the first conjunct. In theoretical linguistics, gapping is a well-studied type of ellipsis (e.g., Jackendoff, 1971; Johnson, 2009; Kuno, 1976; Repp, 2009; Ross, 1970; Sag & Hankamer, 1984; Steedman, 1990). The situation is different, however, for research in psycholinguistics and psychology of language. Studies addressing the processing of elliptical constructions have focused on a particular ellipsis type, i.e., verb-phrase ellipsis (e.g., Katherine ordered a cup of tea and Humphrey did, too). Within this work, different accounts have been proposed as to how the elided element is reconstructed during processing (e.g., Frazier & Clifton, 2001: copying mechanism account; Frazier & Clifton, 2005: shared-structure account; Martin & McElree, 2008, 2010: cue-dependent direct access account). Though the issue as to the reconstruction mechanism is not yet settled, findings from studies on verb-phrase ellipsis and on sluicing, i.e. elision of an entire clause except for a wh-word (e.g., Humphrey ordered something but Spencer doesn't know what Humphrey ordered) converge in suggesting that the elided element is reconstructed in real time — immediately when or shortly after encountering the ellipsis (e.g., Callahan, Walenski, & Love, 2012; Poirier, Wolfinger, Spellman, & Shapiro, 2010; Shapiro & Hestvik, 1995; Shapiro, Hestvik, Lesan, & Garcis, 2003: Yoshida, Dickey, & Sturt, 2013). Verb gapping has received only little attention in empirical studies on the processing of ellipsis. Streb, Henninghausen, and Rösler (2004) contrasted different anaphoric constructions, including verb gapping as an example of surface anaphora (Hankamer & Sag, 1976) and pronouns as an example of deep anaphora. Carlson (2001), Carlson, Dickey, and Kennedy (2005) and Hoeks, Redeker, and Hendriks (2009) addressed the interpretation of sentences such as (2) that are ambiguous between a gapping reading (Katherine met Spencer yesterday and Humphrey met Spencer today) and a “conjunctionreduction” reading (left peripheral deletion: Katherine met Spencer yesterday and Katherine met Humphrey today). 2. Katherine met Spencer yesterday and Humphrey today. There are only a few studies on verb gapping that are directly concerned with the reconstruction of the missing verb information and explored the time course of the reactivation of the gapped verb (Hofmann, 2006; Kaan, Overfelt, Tromp, & Wijnen, 2013; Kaan, Wijnen, & Swaab, 2004). Hofmann (2006) employed a probe-recognition task to investigate whether gapped verb information is reactivated in real time. Participants read gapping sentences such as (3) and corresponding non-elliptical full forms, word-by-word, self-paced. There were two different test points for the probe recognition task; the probe word was either presented after the last word of the first conjunct [TP1] or after the direct object in the second conjunct [TP2]. The probe word was either the direct object noun of the first conjunct (e.g., vegetables) or the finite verb (e.g., sell). 3. Market women sell fresh vegetables to critical customers [TP1] and thieves valuable jewelry [TP2] to wealthy collectors. [(From Hofmann, 2006; translated from German)]

105

For both, the gapping sentences and the corresponding full forms, an interaction between test point and probe word was found. At the first test point, TP1, the probe recognition latencies for the noun and verb probes did not differ. However, at the second test point, TP2, probe recognition latencies for verb probes were shorter than those for noun probes. This specific result pattern and the fact that it was the same for the gapping versions and the full-form versions of the sentences could be considered as evidence that gapped verb information was reactivated online. Yet, there were also differences in the results for the gapping versions and full forms. With gapping sentences, probe recognition latencies for verb probes at TP2 were longer than at TP1, whereas with full forms probe recognition latencies for verb probes were shorter at TP2 than at TP1. Moreover, the finding of shorter latencies with gapping sentences for verb probes compared with noun probes at TP2 may not reflect the reactivation of the gapped verb information. This finding could also be accounted for when assuming that the missing verb was not yet reconstructed when encountering the probe word. According to this interpretation, the effect on the probe words arose after the presentation of the probe word rather than resulting from reactivation of the verb information prior to encountering the probe word (see McKoon & Ratcliff, 1986; see also Nicol, Swinney, Love, & Hald, 2006). Thus, the study by Hofmann (2006) does provide preliminary albeit not unequivocal support for a real-time reactivation of gapped verb information. Kaan et al. (2004) conducted an ERP-study and employed a plausibility manipulation paradigm to explore the time course of processing gapped verbs. More specifically, they investigated whether and when comprehenders notice an implausibility of the object noun in the second conjunct, by juxtaposing two versions of gapping sentences as in (4), differing in the gapped verb and its selectional restrictions on the argument, such that the object noun was either implausible (4a) or plausible (4b). 4a. Jane drank a coffee with cream and Bill a sandwich with cheese. 4b. Jane ordered a coffee with cream and Bill a sandwich with cheese. [(From Kaan et al., 2004)] Event related potentials to the object noun in the second conjunct (e.g., sandwich) in the two conditions revealed an N400 in the implausible condition relative to the plausible condition, followed by a P600. This early effect of implausibility suggests that the gapped verb information was reconstructed in real time, i.e., at the gap or shortly after it. However, the finding is based on specific test conditions: reading of implausible sentences and a fixed presentation rate. Furthermore, the plausibility effects observed by Kaan et al. (2004) could not be replicated in a follow-up study (Kaan et al., 2013). To sum up, findings from a behavioral study and an ERP study are consistent with a real-time reactivation of gapped verb information. However, the evidence is scarce and not clear-cut. One of the goals of the present study was to contribute to this issue by employing a different methodological approach. Rather than using a probe-recognition or plausibility-manipulation paradigm, the present study adopted a paradigm, which has been developed within the embodied-simulation framework, i.e., the reading-by-rotating paradigm introduced by Zwaan and Taylor (2006, Experiment 4; see also Taylor, Lev Ari, & Zwaan, 2008; Taylor & Zwaan, 2008; Zwaan, Taylor, & de Boer, 2010). In the reading-by-rotating paradigm, participants are presented with sentences implying either a clockwise manual rotation (see (5a)) or a counter-clockwise manual rotation (see (5b)). The sentences are presented frame-by-frame (the vertical lines in the example sentences indicate the frame boundaries). Participants advance from one frame to the next by turning a knob either in clockwise or in counter-clockwise direction. 5a. His pencil | was dull | so | before | the | SAT | he | sharpened | his | pencil. 5b. He | realized | that | the | music | was | too loud | so | he | turned down | the | volume. [(From Zwaan & Taylor, 2006)]

106


The main finding of studies employing the reading-by-rotating paradigm is an effect of compatibility between the linguistically implied rotation direction and the direction of the knob turning (Taylor & Zwaan, 2008; Taylor et al., 2008; Zwaan & Taylor, 2006; Zwaan et al., 2010). That is, reading times were affected by whether the knobturning direction matched or mismatched the direction of the implied rotation. Moreover, the effect appears to be immediate and short-lived. In the study by Zwaan and Taylor (2006; Experiment 4), the compatibility effect obtained for the rotation-implying frame (e.g., |sharpened|, |turned down|), with shorter reading times in the match than in the mismatch condition. For any of the other (preceding or subsequent) frames, there was no effect of compatibility on the reading times. Let's now turn back to gapping and the present study and apply the embodied-simulation view and the reading-by-rotating paradigm to the issue of the reactivation of gapped verb information. According to the embodied-simulation view, comprehenders understand an action description by mentally simulating the described action. With regard to gapping sentences, such as (6), this implies that comprehenders simulate not only the action that is indicated by the overt verb in the first conjunct but also the action that is elliptically described in the second conjunct. Thus, effects of compatibility, as for instance obtained with the reading-by-rotating paradigm, are to be expected not only for overt verbs but also for gapped verbs. Over and above, the occurrence of an action-gap compatibility effect would be telling with regard to the issue of the time course of reconstructing the gapped information. In this regard, the reading-by-rotating paradigm seems to be particularly well-suited as it enables time course tracking and has been proven to reveal immediate and short-lived effects. 6. Katherine opens a tube of mustard in her kitchen and Spencer a tube of mayonnaise in his kitchen So, in the present study the reading-by-rotating paradigm was adopted with the aim of investigating whether an effect of compatibility between implied rotation direction and knob turning direction would arise not only for the first conjunct of a gapping sentence (e.g., for the phrase opens a tube of mustard in (6)) but also for the elliptical second conjunct (e.g., the phrase a tube of mayonnaise in (6)). The implications of finding such a compatibility effect for both, the overt verb and the gapped verb, would be twofold. First, with regard to verb gapping, such a finding would provide additional evidence for the real-time reactivation of the elided verb information. In addition, it would indicate that the reading-by-rotating paradigm is a suitable method for studying the processing of verb gapping, thus contributing to methodological variety. Second, with regard to the embodied-simulation view, finding a compatibility effect for gapped verbs would demonstrate that mental simulation is not bound to the processing of overt words. 2. Experiment 2.1. Method 2.1.1. Participants Thirty-six students from Humboldt-Universität zu Berlin participated in the experiment. They gave informed consent to participate in the experiment and received a monetary reimbursement for their participation. 2.1.2. Materials The materials consisted of short narratives texts. There were 24 experimental texts, 52 filler texts, and four practice texts. All texts were three sentences long, and were written in the present tense. The first sentence of each text introduced two characters. The second sentence described actions or events involving the two characters. The third and final sentence completed the narrative. In all experimental texts, the second sentence was a gapping sentence. The gapping

sentences were made up of two clauses, conjoined by the coordinating conjunction und (and). The first conjunct consisted of a subject phrase (proper name), followed by a finite verb, an object noun phrase, and a locative prepositional phrase. The second conjunct had the same structure; however, the finite verb was omitted, resulting in a gapping construction (see sample text in (7)). All employed gapping sentences unambiguously had a gapping reading. An initial non-gapping (conjunction-reduction) reading, when encountering the first two words of the second conjunct, was implausible, as it would result in a semantically anomalous interpretation (e.g., opening/closing Adrian). The locative prepositional phrases in the first and second conjunct were included to allow for the control and measurement of spillover effects. There were two versions of each experimental text. In the counterclockwise version, the gapping sentences described actions that require a counter-clockwise manual rotation (see (7a): opens a lemonade bottle). In the clockwise version, the gapping sentences described actions that require a clockwise manual rotation (see (7b): closes a lemonade bottle). The first and final sentences of the experimental texts also differed in the two versions, i.e., their content was conformed to the respective version of the gapping sentence. 7a. A lot of children arrived at the birthday party and Tina and Adrian are busy providing beverages for everyone. Tina opens a lemonade bottle on the balcony and Adrian a juice bottle in the children's room. Both are happy that so many children are attending the party. (counter-clockwise version) 7b. The birthday party is over and the birthday child's parents, Tina and Adrian, take care of the leftover beverages. Tina closes a lemonade bottle on the balcony and Adrian a juice bottle in the children's room. Both are happy that many children attended the party. (clockwise version) [(Sample experimental text, translated from German)] In the gapping sentences, two different verbs were used in the two versions: öffnen (to open) in the counter-clockwise version, and schließen (to close) in the clockwise version.1 The two verbs were paired with different object nouns referring to objects that are to be closed with a clockwise manual rotation, and to be opened with a counterclockwise manual rotation. The objects were selected on the basis of the results of a pre-test. Thirty-two students of Saarland University were presented with a questionnaire on 138 objects. Of these objects, 106 objects served as target objects, i.e., objects that may be opened or closed by a manual rotation (e.g. mustard tube, jam jar, petrol can). The remaining 32 objects were distractor objects, i.e., objects that are to be opened or closed by a different action (e.g., drawer, wardrobe, baking oven). For each object, participants had to judge how likely it is that it is closed through a clockwise manual rotation or opened through a counter-clockwise manual rotation, respectively. Half of the participants first had to judge all objects with regard to the opening action and were then presented with the same objects with the task to judge the closing action. For the other half of participants, this order was reversed. The objects were presented to the participants in mixed 1 The selection of verbs to be used in the critical gapping sentences was guided by two major constraints. First, the structure of the second conjunct was to be equal to the structure of the first conjunct, except for the gap. In particular, the two conjuncts should not differ in the position of the action-direction conveying frame within the conjunct. Second, the direction of the linguistically implied manual rotation was to be manipulated within items. These constraints rendered many of the German verbs that denote manual-rotation actions inapplicable, namely verbs that are particle verbs (e.g., ein-/aus-schrauben [to screw in/out]), because the action specifying particle would have been detached from the verb, in a clause-final position, i.e. after the prepositional phrase, and verbs that denote manual-rotation actions for which there is no counterpart in the reverse direction (e.g., spitzen [to sharpen]). Thus, with the given constraints it was not possible to have a different verb per item and version, such that repetition of verbs across items could not be avoided. From the few verbs that meet the two constraints, only the two selected verbs, i.e. öffnen (to open) and schließen (to close), are combinable with a multitude of objects and are frequent and common verbs, such that repeating these verbs is feasible as well as not unusual.


orders. For each likelihood judgment, participants first had to imagine that someone would close or open the object, respectively. They then had to rate on a 5-point scale the likelihood that closing the imagined object is done by turning the hand clockwise or that opening the object is done by turning the hand counter-clockwise, respectively. The task was illustrated in the instruction to the questionnaire by drawings of the two hand turning actions on an object depicted from an internal perspective, i.e. viewpoint of someone who is performing the hand turning. The rating scale ranged from −2 (“very unlikely”) to 2 (“very likely”). Of the 106 target objects of the pre-test, 24 objects were selected to be employed in the experimental texts. For each of the selected objects, the distributions of the likelihood ratings in both, the opening and closing action condition were such that the most frequent rating was “very likely” (mode = 2) and that the mean was greater than 1. The names of the 24 selected objects served as the object nouns in the gapping sentences of the 24 experimental texts. Each gapping sentence included two different object nouns, one in the first conjunct and the other in the second conjunct. All selected objects occurred twice, that is, in the gapping sentences of two experimental texts, either as the object noun in the first conjunct or as the object noun in the second conjunct. Each of two experimental texts that shared an object noun in the gapping sentence differed with regard to the second object noun and with regard to setting. The filler texts and practice texts were similar to the experimental texts with respect to topics, style, and length. Twenty-four of the filler texts also included a gapping sentence. Different verbs were used for the filler gapping sentences to avoid that the verbs öffnen (to open) and schließen (to close) would turn to cues for gapping in the course of the experiment. The sentences of all texts (experimental, filler, and practice) were divided into frames at intuitively natural boundaries, consisting of one to seven words. The gapping sentences of the experimental texts were all divided into six frames according to the following schema (see Table 1): Frame 1: proper name of character A; Frame 2: finite verb (closes/opens) + object noun phrase (overt-verb frame); Frame 3: prepositional phrase; Frame 4: conjunction (and) + proper name of character B; Frame 5: (gap) object noun phrase (gapped-verb frame); Frame 6: prepositional phrase. To encourage the participants to read the texts carefully, 28 texts (all fillers) were followed by a yes/no comprehension question pertaining to the content of the given text. 2.1.3. Apparatus The input device, which enabled self-paced frame-by-frame text presentation, was a knob that was mounted on a box, such that it allowed a manual hand rotation in the horizontal plane. The knob was 4.5 cm in diameter. It contained springs, such that it returned to its center position when being released. A knob turning of approximately 60° triggered the presentation of the next frame and the logging of the reading time for the given frame. 2.1.4. Design and procedure Direction of knob turning and direction of linguistically implied rotation were manipulated within-subject. Participants were randomly assigned to four groups, and the experimental texts were randomly Table 1 Illustration of the frame-by-frame presentation for a sample gapping sentence. First conjunct Frame 1 Frame 2 (overt verb) Frame 3

Second conjunct Tina opens/closes a lemonade bottle on the balcony

Frame 4 Frame 5 (gapped verb) Frame 6

and Adrian a juice bottle in the children's room

Note. Each gapping sentence was presented to the participants either in the counterclockwise version (opens) or in the clockwise version (closes).

107

assigned to four sets. The four conditions were assigned to the groups and sets according to a 4 (group) × 4 (set) × 4 (condition) Latin square. Each participant was initially randomly assigned to the clockwise knob turning or counter-clockwise knob turning condition for the first block of the experiment. Halfway through the experiment, they were instructed to change the knob turning direction for the second block of the experiment. Experimental and filler texts were presented to the participants in mixed random orders. Each sequence of each of the two blocks started with four filler texts serving as warm-ups. The texts were presented frame by frame on a computer screen in 20 pt Arial font. The participants advanced through each text by turning a knob either in clockwise or in counter-clockwise direction, depending on the respective block condition. The time interval between the presentation of a frame on the screen and participants' subsequent knob turning was defined as the reading time for the respective frame. After each knob turning, participants had to let the knob turn back to its center position. They were asked to keep their hand on the knob during a whole trial, that is, until they had read all frames of a given text. Comprehension questions were announced by three question marks. Participants responded by pressing either a yes-button or a no-button. Participants were tested individually. They were instructed to read all texts carefully at their normal reading pace, and to be prepared to answer questions about the texts. In each of the two blocks, the procedure was illustrated by means of two practice trials. 2.2. Results Analyses were performed on the reading times per frame of the gapping sentences and were based on residual reading times, adjusted for frame length in terms of number of characters. Data from one participant were excluded from all analyses because the accuracy of this participant's responses to the comprehension questions was not significantly better than chance. Reading times for each of the remaining participants were regressed on the number of characters per frame. Prior to the regression analyses, all reading times shorter than 100 ms were discarded. Separately for each participant, a linear regression equation was computed, predicting frame reading time from the number of characters, by using the frames of the second sentences of the filler texts. On the basis of the regression equations, the readingtime residuals for the frames of the critical gapping sentences were determined. Residual reading times shorter than −600 were excluded. Outliers in the remaining data were determined separately for the six sentence frames. Residual reading times deviating more than 2.5 standard deviations from the respective condition mean were eliminated. Altogether, 3.80% of the data were discarded from the analyses. The mean residual reading times for the six frames of the gapping sentences in the two match conditions are displayed in Fig. 1. The data were submitted to a 6 (sentence frame) by 2 (match) analysis of variance. The analysis yielded a significant main effect of sentence frame, F 1 (5,170) = 9.74, MSE = 9361, p b .001; F 2 (5,115) = 17.20, MSE = 3799, p b .001, which is of little interest however, as the sentence frames were not matched for complexity and differed in position within the sentence. The main effect of match was marginally significant, F1(1,34) = 3.19, MSE = 6122, p = .083; F2(1,23) = 3.66, MSE = 3965, p = .068. This was qualified by a significant Frame × Match interaction, F1(5,170) = 2.77, MSE = 2999, p = .020; F2(5,115) = 2.41, MSE = 2196, p = .041. To unpack the interaction, pairwise comparisons (two-tailed t-tests) were made between the match and mismatch condition separately for each of the six sentence frames. These revealed a match effect for the overt-verb frame in the first conjunct, which was significant in the analysis by participants and narrowly failed to reach significance in the analysis by items (Frame 2: t1(34) = 2.34, p = .026; t2(23) = 2.01, p = .056) and a significant match effect for the gapped-verb frame in the second conjunct (Frame 5: t1(34) = 2.12, p = .041; t2

108


match mismatch

Frame 1 Tina

Frame 2 closes/opens a lemonade bottle

Frame 3

Frame 4

Frame 5

on the balcony

and Adrian

a juice bottle

Frame 6 in the children’s room

Fig. 1. Mean residual reading times (in ms) per frame of the gapping sentences in the match and mismatch condition. Error bars represent within-subject 95% confidence intervals (see Masson & Loftus, 2003) associated with the contrast between the match and mismatch condition in the respective frame.

(23) = 2.68, p = .014). For both frames, reading times were longer in the match condition than in the mismatch condition (see Fig. 1). There was no significant effect of match for any of the other frames (Frame 1: t1(34) = 0.85, p = .401; t2(23) = 0.53, p = .601; Frame 3: t1 (34) = 0.19, p = .854; t 2 (23) = 0.47, p = .640; Frame 4: t 1 (34) = 0.72, p = .477; t2(23) = 0.94, p = .358; Frame 6: t1(34) = 0.99, p = .325; t2(23) = 0.95, p = .354). Thus, the interaction of sentence frame and match seems to be driven by the presence of a match effect in the two action-conveying frames and its absence in the other frames. To check for potential differences between the overt-verb frame and the gapped-verb frame, the reading-time residuals for the two frames were submitted to a 2 (overt verb/gapped verb) by 2 (match/mismatch) analysis of variance. This demonstrated significant main effects of verb frame (F1(1,34) = 8.79, MSE = 14,533, p = .005; F2(1,23) = 17.29, MSE = 5294, p b .001) and of match (F1(1,34) = 8.12, MSE = 6723, p = .007; F2(1,23) = 9.57, MSE = 3776, p = .005) but no interaction effect (F1(1,34) = 0.10, MSE = 4252, p = .755; F2 (1,23) = 0.02, MSE = 3404, p = .880), indicating that the match effect was not different for the overt and gapped verb frame. Summing up the results, a significant compatibility effect was found for the overt-verb frame in the first conjunct (e.g., |closes/opens a lemonade bottle|) as well as for the gapped-verb frame (e.g., |a juice bottle|) in the second conjunct. There was no significant compatibility effect for any of the other sentence frames. However, the direction of the compatibility effect for the two action-implying frames was opposite to the effect that has been found in several studies by Zwaan and Taylor and colleagues (Taylor & Zwaan, 2008; Taylor et al., 2008; Zwaan & Taylor, 2006; Zwaan et al., 2010). In the present experiment, reading times in the two action-conveying frames were found to be longer – rather than shorter – when there was a match between the direction of the linguistically conveyed action and the actual knob rotation compared with when there was a mismatch. I get back to this issue in the Discussion section. With regard to the specific aim of the experiment, the crucial finding is (1) that the pattern of results for the second, elliptical conjunct equals the pattern of result for the first conjunct, and (2) that the effect of compatibility was specific to the two actionimplying frames. On the one hand, the results indicate that languageinduced motor activation is not bound to the processing of overt words

but also obtains for elliptical constructions. On the other hand, the result for the gapped-verb frame suggests that the missing verb information was reactivated in real time, i.e., during processing of the frame that revealed the gap. 3. Discussion The goal of the present study was to explore if there would be an interaction between language comprehension and motor action not only for overt action verbs but also for gapped verbs. In a reading-byrotating paradigm experiment, an effect of compatibility between the directions of a linguistically implied and an actually performed action was found not only for overt verbs but also for verb gaps. As outlined in the introduction, this finding is revealing both for research on verb gapping processing and for research within the embodied-simulation view. Before looking into the twofold implications of the present finding, however, it seems to be indicated to go into the matter of the direction of the compatibility effect in the present experiment. The present result differs from the common finding from experiments employing the reading-by-rotating paradigm. In the present study, the compatibility effect is due to shorter reading times in the mismatch condition than in the match condition, whereas in previous studies, it was due to shorter reading times in the match condition than in the mismatch condition (Taylor & Zwaan, 2008; Taylor et al., 2008; Zwaan & Taylor, 2006; Zwaan et al., 2010). As a matter of fact, when considering studies on the interaction between language comprehension and actual action beyond the reading-by-rotating paradigm, the results are mixed with regard to the type of the effect. Some studies have found a match advantage/mismatch disadvantage effect, with shorter latencies when there was a match between the linguistically implied action and the actually to be performed action compared with when there was a mismatch (e.g., Bergen & Wheeler, 2005, 2010; Bub & Masson, 2010; Glenberg & Kaschak, 2002; Masson, Bub, & Newton-Taylor, 2008). Yet, there are also several studies that yielded a mismatch advantage/match disadvantage effect as in the present experiment, with shorter latencies in the mismatch condition compared with the match condition (de Vega et al., 2013; de Vega & Urrutia, 2011; Buccino et al., 2005; see also Borreggine & Kaschak, 2006; Boulenger et al.,


2006; Nazir et al., 2008). It is not yet conclusively settled how the mixed results, i.e., the different effect types, can be reconciled. In fact, the embodied-simulation view alone does not allow deriving specific predictions as to which factors affect the direction of the compatibility effect. Moreover, the nature of compatibility effects is still not clear, for instance whether a match advantage/mismatch disadvantage effect is due to facilitation in the match condition, or inhibition in the mismatch condition, or both. Research on these issues is still in its beginning. According to some proposals, temporal parameters may be crucial (Bergen, 2007; Borreggine & Kaschak, 2006; Boulenger et al., 2006; de Vega et al., 2013; de Vega & Urrutia, 2011; Nazir et al., 2008; see also Kaschak, Zwaan, Aveyard, & Yaxley, 2006; Meteyard et al., 2007; see also Bub & Masson, 2012), i.e., the type of effect may depend on the temporal overlap between mentally simulating described actions and planning actually to be performed motor actions. I will come back to the role of temporal parameters in the following, when I will consider three possible accounts of the opposite types of effect in the present experiment and the preceding studies employing the reading-by-rotating paradigm (Taylor & Zwaan, 2008; Taylor et al., 2008; Zwaan & Taylor, 2006; Zwaan et al., 2010). Though in the present experiment, the same paradigm was employed as in the studies by Zwaan, Taylor and colleagues, there were still differences in the particular experimental task demands. One difference pertains to the specific knob-turning procedure. In the experiments by Zwaan, Taylor and colleagues, participants had to turn the knob by approximately five degrees to trigger the presentation of the next frame. The knob contained springs, such that it would return to its center position when being released. Participants were instructed to keep hold of the knob until the final sentence frame; only then the knob was to be released to the center position. In the present experiment, participants had to turn the knob for approximately 60°. The knob likewise contained springs. However, participants were not instructed to keep hold of the knob; rather, after each knob turning, they had to let the knob turn back to its center position. They were asked to keep their hand on the knob during each whole trial, that is, during processing all the frames of a given text. Though it cannot be excluded that the opposite types of effect are simply due to the difference in the degree of knob turning, this does not appear to offer a compelling account. But the difference with regard to the instruction on keeping hold of versus releasing the knob may have been crucial and opens up two possible accounts. Different procedures – same type of effect? The finding of a match advantage/mismatch advantage effect in the studies by Zwaan, Taylor and colleagues (Taylor & Zwaan, 2008; Taylor et al., 2008; Zwaan & Taylor, 2006; Zwaan et al., 2010) is commonly interpreted as an effect of language comprehension on action, i.e., it is assumed that mentally simulating the described action in the match condition brings about a preactivation of the motor program required for the knob turning. Yet, this is not the only possible interpretation; the finding may instead be due to an effect of action on language comprehension. Suppose this would be true, that is, suppose that the finding reflects an effect of the preceding knob turning on the processing of the respective current frame. Then, the prediction would be that the reading times for a rotation-implying frame would depend on the preceding knob turning, such that the reading times should be shorter if the direction of the preceding knob turning matches the direction of the implied rotation in the current frame. Now, consider that in the present experiment – different from the studies by Zwaan, Taylor, and colleagues –, participants had to release the knob to its center position after each turning, with their hand still on the knob — resulting in a back turning of the knob. Then, the knob turning immediately preceding each frame was this back turning of the knob. As a consequence, the direction of the knob turning prior to processing the current frame (i.e., the back turning) was opposite to the direction according to the experimental specified knob condition. For example, when participants had to turn the knob in counter-clockwise direction, the direction of

109

the back turning was clockwise. Thus, when assuming that it is the immediately preceding actual rotation that causes the effect, which in the present experiment was the turning back of the knob, then this reverses the assignment of conditions to the levels match and mismatch. Match and mismatch according to knob direction condition are mismatch and match, respectively, according to back turning direction. Along these lines of reasoning, the match disadvantage/mismatch advantage effect would have to be considered as a match advantage/ mismatch disadvantage effect. According to this back-turning account, the type of effect in the present experiment would not be opposite to the type of effect observed by Zwaan, Taylor, and colleagues, but rather would be of the same type. There is, however, an important caveat to the back-turning account, namely that the back turning of the knob was caused by the springs, which returned it to its center position. Thus, the back turning did not involve the planning and control of an active manual rotation. Accordingly, any potential effect of the knob's back turning on processing the linguistic information of the subsequent frame cannot be due to the programming of a to be executed motor plan. One loophole would be to assume an effect in terms of an effect of motion observation, i.e. observing the knob turning back to its center position. However, this is rendered unlikely by findings from research on the interaction between action observation and action performance, which indicate that effects of action observation are bound to actions that are believed to originate from a biological agent (e.g., Kilner, Paulignan, & Blakemore, 2003; Stanley, Gowen, & Miall, 2007). Another loophole would be to attribute the hypothetical effect of the back turning to an effect of feature overlap of a passive movement and a described motion, i.e., an effect of compatibility effect between the direction of being back-rotated and the linguistically implied manual rotation. This second loophole also appears questionable considering that the encoding of a passive motion crucially differs from the encoding of a self-motion (e.g., Wutte, Glasauer, Jahn, & Flanagin, 2012). Aside from the specific problems of the two loopholes, the principal objection against the back-turning account is that it has to exclude motor planning as a basis for explanation. It thereby conflicts with prevailing accounts of compatibility effects between language and action as well as with evidence from neuroimaging studies that processing action descriptions leads to activation in motor brain areas that are involved in action planning and controlling (e.g., Aziz-Zadeh, Wilson, Rizzolatti, & Iacoboni, 2006; Hauk, Johnsrude, & Pulvermüller, 2004; Tettamanti et al., 2005). Different procedures — different types of effect? According to the second account to be considered here, the reversed effect in the present experiment compared with the findings by Zwaan, Taylor, and colleagues actually reflects a different type of effect. The second account acts on the assumption that the reading-by-rotating compatibility effect (either match advantage/mismatch disadvantage or mismatch advantage/ match disadvantage) is due to an effect of language comprehension on action, i.e., an effect of processing of the current frame on planning the subsequent knob turning. In addition, it supposes that temporal parameters are crucial, i.e., the temporal overlap between mentally simulating a linguistically implied rotation and planning an actual manual rotation. Now reconsider the different knob-turning procedures: in the present experiment, participants had to release the knob after each turning; in the experiments by Zwaan, Taylor, and colleagues, participants had to keep holding the knob. This difference in knobturning procedure may bring about a difference with regard to temporal overlap. For the present experiment, it seems conceivable that participants initiated the planning of the knob turning only after simulating what was described in the current frame. In contrast, with the procedure employed by Zwaan, Taylor, and colleagues, it seems conceivable that participants initiated the planning of the knob turning while simulating what was described in the current frame. Research on action and perception indicates that this difference may be crucial (Hommel, Müsseler, Aschersleben, & Prinz, 2001: theory of event

110


coding (TEC); see Borreggine & Kaschak, 2006, for an application of the TEC to the action–sentence compatibility effect). In case of a match between the direction of a linguistically implied action and an actually to be performed action, temporal overlap of action simulation and action planning may lead to facilitation, whereas temporal succession may lead to inhibition. This account seems to be in line with the results of the study by Borreggine and Kaschak (2006), in which an effect of action–sentence compatibility in terms of a match advantage/mismatch disadvantage effect was obtained when the direction of the response action was specified at sentence onset, such that participants could plan the action during sentence processing, but was not obtained when the direction was specified after sentence presentation. However, the account seems to be inconsistent with the results of a study by de Vega et al. (2013) which are opposite to the findings by Borreggine and Kaschak. De Vega et al. found a mismatch advantage/ match disadvantage effect when the processing of an action word overlapped with a motor task, but not when there was a delay between action word processing and the motor task. A possible explanation for the mixed results may lie in the fact that the study by de Vega et al. used a dual-task paradigm, such that the temporal overlap did involve action planning and execution. One vs. two main characters — different perspectives? The third account of the opposite direction of the compatibility effect in the present experiments and in the experiments by Zwaan, Taylor, and colleagues pertains to a difference with regard to the materials. Whereas in the experiments by Zwaan, Taylor, and colleagues, each of the employed sentences involved only one main character, there were two main characters in each text of the present experiment. This difference may have been crucial insofar as it might involve a difference with regard to perspective adoption. It is empirically well established that comprehenders take over the protagonist's perspective for sentences or texts with a unique main character (e.g., Black, Turner, & Bower, 1979; Borghi, Glenberg, & Kaschak, 2004; Bryant, Tversky, & Franklin, 1992; Horton & Rapp, 2003; Sanford, Clegg, & Majid, 1998). Findings from a study by Franklin, Tversky, and Coon (1992) on spatial mental models suggest that matters may be different when there is more than one relevant character; for spatial descriptions involving two observers, they found evidence for the adoption of a neutral perspective rather than two internal, i.e., character-centered, perspectives. From this, it could be conjectured that the participants of the experiments by Zwaan, Taylor, and colleagues adopted the internal perspective of the single main character, whereas the participants of the present experiment did not adopt two internal perspectives, one for each of the two main characters, but rather took over an external, observer perspective. Drawing from a finding of a study by Sato, Sakai, Wu, and Bergen (2012; see also Brunyé, Ditman, Mahoney, Augustyn, & Taylor, 2009; Sato & Bergen, 2013) on perspective modulation, this external perspective may have been such that participants simulated observing the actions of the two main characters from an “en face” view point. Thus, according to this account the opposite directions of effect in the present experiment and in the experiments by Zwaan, Taylor, and colleagues are due to a difference in adopted perspective: “en face” perspective for the two main characters in the present experiment versus internal perspective for the single main character in the experiments by Zwaan, Taylor, and colleagues. To sum up, the first and second account of the reversed effect direction in the present experiment compared with the findings by Zwaan, Taylor, and colleagues pertain to differences in the specific knob-turning procedure (Account 1: direction of knob turning preceding the respective current frame; Account 2: temporal overlap of action simulation and action planning) and the third account attributes the different effect directions to a difference in the materials (Account 3: single versus two main characters and perspective adoption). All three accounts link up to important and yet not fully understood issues within the embodied-simulation framework. As discussed above, there is a serious caveat to Account 1. Hence,

Account 1 can be dismissed. An issue with Account 2 is that there is a seemingly conflicting finding. Yet, it may be premature to dismiss Account 2, considering that the finding has been obtained with a paradigm to which Account 2 may not be applicable. Regarding Account 3, there is neither a conflicting finding nor a direct evidence in favor of it. It remains an empirical task to evaluate the scope and value of Accounts 2 and 3. Turning back to the specific purpose of the present experiment, i.e., investigating the processing of verb gapping, the results can be considered to show a clear pattern. The results for the three frames of the elliptical second conjunct were essentially the same as the results for the three frames for the first (non-elliptical) conjunct. For both clauses, a compatibility effect was found only for the actionimplying frame, which contained either an overt verb or a verb gap — with the effect being of the same type for both, the overt-verb frame and the gapped-verb frame. No compatibility effect was found for any of the preceding or subsequent frames — indicating that the effect of compatibility was specific to the two actionimplying frames. The fact that the compatibility effect obtained not only for the overt-verb frame but also for the gapped-verb frame indicates that participants reactivated the missing verb information in real time. The present finding thus contributes to research on verb gapping processing. It adds to existing results that are in line with a realtime reconstruction of gapped verbs (Hofmann, 2006; Kaan et al., 2004) with a novel method. The most straightforward interpretation of the finding of a compatibility effect for the gapped- verb frame seems to be, that participants reactivated the missing verb during processing that frame. Yet, the sustained reactivation hypothesis proposed by Callahan, Shapiro, and Love (2010) suggests an alternative interpretation. The sustained reactivation hypothesis relates to parallelism in coordinate structures. With regard to the processing of conjoined sentences, it maintains that a cue for parallelism such as the conjunction and triggers the reactivation of elements from the first conjunct. Consistent with this hypothesis, Callahan et al. (2010) found evidence for the reactivation of the verb from the first conjunct at a test point [TP2] immediately following the conjunction and as well as at test points [TP3 and 4] within the second conjunct, whereas for an initial test point [TP1] within the first conjunct (some words after the verb position) no activation of the verb was found (see (8) for an example sentence). 8. The doctor read the chart of the child with the broken arm [TP1] during his morning rounds, and [TP2] the insurance agent in [TP3] the tacky suit did as well [TP4] in order to become more familiar with the case. [(From Callahan et al., 2010)] Applying the sustained reactivation hypothesis to the present experiment, suggests that the compatibility effect for the gapped-verb frame does not reflect the reactivation of the missing verb information at or shortly after the gap but rather reflects a prior, anticipatory reactivation of the verb of the first conjunct, supposedly triggered by the conjunction und (and) when processing the preceding, first frame of the second conjunct. Yet, there is reason to doubt this alternative interpretation of the present results. First, differently from what could be predicted from the sustained reactivation hypothesis, no compatibility effect was found for the first frame of the second conjunct that included the conjunction, i.e., the supposed cue for parallelism. Admittedly, the lack of an effect for the frame including the conjunction does not preclude the possibility that processing the conjunction elicited the reactivation of the verb of the first conjunct. The verbs employed in the experimental items (öffnen [open] and schließen [close]) are “underspecified” in that the specific action involved in opening and closing is indeterminate without context and depends on the particular object (e.g., opening/ closing a jar versus a door). This even holds, when assuming that the


first conjunct constrains the possible objects, for example to bottles (consider bottles with twist caps versus bottles with corks or with crown caps). Yet, the lack of an effect for the conjunction-including first frame of the second conjunct does at least demonstrate that the supposed reactivation of elements from the first conjunct does not involve action-specific information. Second, the finding by Callahan et al. (2010) of an anticipatory reactivation of the verb immediately following the conjunction and may not be due to the conjunction serving as a cue for parallelism but rather as a cue for conjunction reduction (e.g. The doctor read the chart of the child […] during his morning round and read the chart of the man with the broken leg). Third, the findings by Callahan et al. (2010) stem from naming latencies in a dual task; while listening to sentences, participants had to name visually presented probe words (inflected verbs that were either semantically related or unrelated to the verb of the first conjunct). Considering these task demands, the observed effects may not reflect a reactivation of the verb triggered by the conjunction, but may have resulted from processes of integrating the probed verbs into the continuing sentence. Fourth, results from an ERP-study on gapping by Kaan et al. (2013) are equivocal with regard to the sustained reactivation hypothesis. When assuming that comprehenders do not or not consistently reactivate the verb information from a first conjunct already when encountering a conjunction but only when encountering an ellipsis, then the earliest possible cue for the verb gap in the materials employed in the present experiment was the determiner of the object DP. Because the determiner and the object noun were presented in one frame, the results are not informative with regard to whether the gap was detected and the missing verb was reactivated already when processing the determiner or later, when processing the noun (see the sample gapping sentence, here repeated as (9)). 9. Tina | opens/closes a lemonade bottle | on the balcony | and Adrian | a juice bottle | in the children's room. Even though the present results leave open the issue as to when exactly during processing the object DP the verb information was reactivated, they may suggest an early integration of the reactivated verb information. The compatibility effect for the gapped-verb frame (|a juice bottle| in (9)) may be interpreted as an indication that the verb was fully integrated, syntactically and semantically, during processing the object DP. This interpretation is consistent with most accounts of reconstruction of elided material (e.g., Frazier & Clifton, 2001: copying mechanism account; Frazier & Clifton, 2005: shared-structure account; Martin & McElree, 2008, 2010: cuedependent direct access account). It attributes the compatibility effect to processes subsequent to the reconstruction and integration of the missing verb information. A different interpretation emerges when assuming that the reactivated information is not (or not solely) a linguistic representation, e.g., at the level of logical form, but rather a non-linguistic, conceptual representation (see Garnham, 2001: mental models account). According to this view, the compatibility effect may be considered as a direct result of reactivating an action representation. From a linguistic perspective, however, a purely conceptual view on reconstruction of elliptical material is dubious. There are several arguments against such a view (e.g., Frazier & Clifton, 2005), particularly with regard to gapping, which is syntactically more restricted than other types of ellipsis. The present experiment employed a paradigm – reading-byrotating – that was developed by Zwaan and Taylor (2006) within the embodied-simulation framework. Here, the reading-by-rotating was applied to study a genuine linguistic phenomenon, verb gapping. The present study provides a preliminary demonstration that this paradigm can be of value in investigating the reactivation of gapped verbs. Thus, the paradigm may contribute to methodological

111

variety in research in the area of ellipsis processing by providing a method that does not involve the interruption of sentence presentation by probe words, nor a fixed presentation rate, nor the presentation of semantically anomalous sentences. However, it may be limited with regard to fine-grained temporal resolution and poses restrictions on the materials (descriptions of particular actions). Finally, I will shortly address what the results of the present experiment imply with regard to the embodied-simulation view per se. If accepting that the compatibility effect for the action-conveying frames in the first and second conjuncts is to be attributed to a mental simulation of the implied actions, then the fact that the compatibility effect did occur not only with the overt-verb frame but also with the gapped-verb frame, can be considered as evidence that the mental simulation of linguistically conveyed actions is not bound to the processing of overt verbs. The present experiment thereby adds to previous studies with the reading-by-rotating paradigm that observed motor resonance effects for action-modifying adverbs (Taylor & Zwaan, 2008, Experiment 1, e.g., On the shelf, he found a closed jar which he opened rapidly) and for adjectives, that constrained the direction of a before mentioned action (Taylor et al., 2008, e.g., He approached the stereo and adjusted the volume. The music was too loud). Findings such as these may indicate that mental simulations are the product of a composition process (see Kaup, Lüdtke, & Steiner, 2012; see Gennari, MacDonald, Postle, & Seidenberg, 2007 and Moody & Gennari, 2010 for neuroscientific evidence of combinatorial processes) rather than a rudimentary byproduct of pure word-based processes. Acknowledgments My thanks go to Carsten Schliewe for building the knob that was employed in the experiment, to Yasmin Dalati, Julia Lengyel, Matthias Lickfett, and Anne Schmidt for their assistance in collecting the data, to Katharina Vnoucek for her help in translating all critical gapping sentences into English, and to Peter Pollmanns and three anonymous reviewers for their valuable comments. Appendix English translation and German version (in parentheses) of the critical gapping sentences of the experimental texts, in counterclockwise version (“opens”) and in clockwise version (“closes”), preceded by an outline of the respective text's setting. Setting: child's birthday party | Tina opens/closes a lemonade bottle on the balcony and Adrian a juice bottle in the children's room. (Tina öffnet/schließt eine Fantaflasche auf dem Balkon und Adrian eine Saftflasche im Kinderzimmer.) Setting: working at a takeaway | Volker opens/closes a mustard tube at the bar table and Jochen a tube of remoulade at the counter. (Volker öffnet/schließt eine Senftube am Stehtisch und Jochen eine Remouladentube am Verkaufstresen.) Setting: newborn nursery | Monika opens/closes a feeding bottle at the sink and Lisa a cough syrup bottle at the medicine chest. (Monika öffnet/schließt ein Milchfläschchen am Spülbecken und Lisa eine Hustensaftflasche am Medikamentenschrank.) Setting: father and son, repairing car and bike | Georg opens/closes a petrol can at the workbench and Philipp a tyre valve on his racing bike. (Georg öffnet/schließt einen Benzinkanister an der Werkbank und Philipp ein Reifenventil an seinem Rennrad.) Setting: baking a cake | Jessica opens/closes a bottle of Amaretto at the beverage shelf and Jonas a jar of peaches next to the juicer. (Jessica öffnet/schließt eine Amarettoflasche am Getränkeregal und Jonas ein Pfirsichglas neben dem Entsafter.) Setting: at a lido | Andi opens/closes a coke bottle next to the cool bag and Steffen a lemonade bottle on the bath towel. (Andi öffnet/ schließt eine Colaflasche neben der Kühltasche und Steffen eine Fantaflasche auf dem Badetuch.)

112


Setting: barbecue party | Achim opens/closes a bottle of apple juice at the beverage table and Carolin a tube of mayonnaise at the buffet table. (Achim öffnet/schließt eine Apfelsaftflasche am Getränketisch und Carolin eine Mayonnaisetube am Büffet.) Setting: feeding baby and preparing dessert | Antje opens/closes a baby food jar on the kitchen counter and Konstantin a jar of lingonberries on the worktop. (Antje öffnet/schließt ein Babybreiglas auf dem Küchentresen und Konstantin ein Preiselbeerglas auf der Arbeitsplatte.) Setting: roommates sharing bathroom | Marianne opens/closes a tube of cream in front of the mirror and Sarah a tube of toothpaste in front of the bathroom cabinet. (Marianne öffnet/schließt eine Cremetube vor dem Spiegel und Sarah eine Zahnpastatube vor dem Badezimmerschrank.) Setting: making potato salad | Helga opens/closes a tube of remoulade in front of the kitchen rack and Stefanie a bottle of vinegar next to the larder cupboard. (Helga öffnet/schließt eine Remouladentube vor dem Küchenregal und Stefanie eine Essigflasche neben dem Vorratsschrank.) Setting: housewarming party | Leonie opens/closes a bottle of Campari in the living room and Saskia a bottle of Amaretto at the house bar. (Leonie öffnet/schließt eine Campariflasche im Wohnzimmer und Saskia eine Amarettoflasche an der Hausbar.) Setting: working at a petrol station | Christian opens/closes a petrol cap on a car and Martin a petrol can next to the car wash facility. (Christian öffnet/schließt einen Tankdeckel an einem Auto und Martin einen Benzinkanister neben der Waschanlage.) Setting: mother and son, preparing his favorite dish | Sven opens/ closes a yoghurt jar next to the fridge and Marion a jar of peaches next to the waffle iron. (Sven öffnet/schließt ein Joghurtglas neben dem Kühlschrank und Marion ein Pfirsichglas neben dem Waffelautomat.) Setting: first day on a houseboat in need of renovation | Grit opens/ closes a tube of toothpaste at the sink and Dirk a bottle of turpentine next to the material cabinet. (Grit öffnet/schließt eine Zahnpastatube über dem Waschbecken und Dirk eine Terpentinflasche neben dem Materialschrank.) Setting: break during bicycle tour | Paul opens/closes a tyre valve on his bike and Heinz a thermos bottle on his carrier. (Paul öffnet/schließt ein Reifenventil an seinem Fahrrad und Heinz eine Thermosflasche auf seinem Gepäckträger.) Setting: making pasta salad | Nina opens/closes a bottle of vinegar at the kitchen table and Carlo a gherkin jar at the sink unit. (Nina öffnet/ schließt eine Essigflasche am Küchentisch und Carlo ein Gurkenglas am Spülschrank.) Setting: bartenders at work | Thomas opens/closes a bottle of juice at the sink and Pascal a bottle of Campari at the bar. (Thomas öffnet/ schließt eine Saftflasche über der Spüle und Pascal eine Campariflasche an der Theke.) Setting: two mothers and their babies having a picnic | Maria opens/ closes a baby food jar on the park bench and Hannah a feeding bottle on the picnic rug. (Maria öffnet/schließt ein Babybreiglas auf der Parkbank und Hannah ein Milchfläschchen auf der Picknickdecke.) Setting: home-made fruit juice and stewed fruit | Matthias opens/closes a bottle of apple juice next to the bottle stand and Sandra a jar of lingonberries next to the freezer. (Matthias öffnet/ schließt eine Apfelsaftflasche neben dem Flaschenständer und Sandra ein Preiselbeerglas neben dem Gefrierschrank.) Setting: at the studio of an art college | Marc opens/closes a thermos bottle in the kitchenette and Astrid a bottle of turpentine at the drawing board. (Marc öffnet/schließt eine Thermosflasche in der Küchenecke und Astrid eine Terpentinflasche am Maltisch.) Setting: working at a vintage car rally | Otto opens/closes a petrol cap of a classic car and Horst a coke bottle on the folding table. (Otto öffnet/schließt einen Tankdeckel an einem Oldtimer und Horst eine Colaflasche auf dem Klapptisch.)

Setting: preparing a cold buffet | Sabine opens/closes a tube of mayonnaise next to the salad bowl and Maike a yoghurt jar next to the blender. (Sabine öffnet/schließt eine Mayonnaisetube neben der Salatschüssel und Maike ein Joghurtglas neben dem Mixer.) Setting: work day at a doctor's surgery | Thorsten opens/closes a cream tube at the examination table and Sylvia a cough syrup bottle in the medication room. (Thorsten öffnet/schließt eine Cremetube an der Krankenliege und Sylvia eine Hustensaftflasche im Medikamentenraum.) Setting: making hot dogs | Anne opens/closes a gherkin jar on the worktop and Markus a mustard tube next to the bread bin. (Anne öffnet/schließt ein Gurkenglas auf der Ablage und Markus eine Senftube neben dem Brotkasten.) References Aziz-Zadeh, L., Wilson, S. M., Rizzolatti, G., & Iacoboni, M. (2006). Congruent embodied representations for visually presented actions and linguistic phrases describing actions. Current Biology, 16, 1–6. Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and Brain Sciences, 22, 577–609. Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617–645. Bergen, B. K. (2007). Experimental methods for simulation semantics. In M. GonzalesMarquez, I. Mittelberg, S. Coulson, & M. J. Spivey (Eds.), Methods in cognitive linguistics (pp. 277–301). Amsterdam: John Benjamins. Bergen, B. K., Lindsay, S., Matlock, T., & Narayanan, S. (2007). Spatial and linguistic aspects of visual imagery in sentence comprehension. Cognitive Science, 31, 733–764. Bergen, B. K., & Wheeler, K. B. (2005). Sentence understanding engages motor processes. In B. G. Bara, L. Barsalou, & M. Bucciarelli (Eds.), Proceedings of the XXVII Annual Conference of the Cognitive Science Society (pp. 238–243). Mahwah, NJ: Lawrence Erlbaum. Bergen, B., & Wheeler, K. (2010). Grammatical aspect and mental simulation. Brain & Language, 112, 150–158. Black, J. B., Turner, T. J., & Bower, G. H. (1979). Point of view in narrative comprehension, memory, and production. Journal of Verbal Learning and Verbal Behavior, 18, 187–198. Borghi, A.M., Glenberg, A.M., & Kaschak, M. P. (2004). Putting words in perspective. Memory & Cognition, 32, 863–873. Borreggine, K. L., & Kaschak, M. P. (2006). The action–sentence compatibility effect: It's all in the timing. Cognitive Science, 30, 1097–1112. Boulenger, V., Roy, A.C., Paulignan, Y., Deprez, V., Jeannerod, M., & Nazir, T. A. (2006). Cross-talk between language processes and overt motor behavior in the first 200 msec of processing. Journal of Cognitive Neuroscience, 18, 1607–1615. Brunyé, T. T., Ditman, T., Mahoney, C. R., Augustyn, J. S., & Taylor, H. A. (2009). When you and I share perspectives: Pronouns modulate perspective taking during narrative comprehension. Psychological Science, 20, 27–32. Bryant, D. J., Tversky, B., & Franklin, N. (1992). Internal and external spatial frameworks for representing described scenes. Journal of Memory and Language, 31, 74–98. Bub, D. N., & Masson, M. E. J. (2010). On the nature of hand-action representations evoked during written sentence comprehension. Cognition, 116, 394–408. Bub, D. N., & Masson, M. E. J. (2012). On the dynamics of action representations evoked by names of manipulable objects. Journal of Experimental Psychology: General, 141, 502–517. Buccino, G., Riggio, L., Melli, G., Binkofski, F., Gallese, V., & Rizzolatti, G. (2005). Listening to action-related sentences modulates the activity of the motor system: A combined TMS and behavioral study. Cognitive Brain Research, 24, 355–363. Callahan, S. M., Shapiro, L. P., & Love, T. (2010). Parallelism effects and verb activation: The sustained reactivation hypothesis. Journal of Psycholinguistic Research, 39, 101–118. Callahan, S. M., Walenski, M., & Love, T. (2012). The processing and interpretation of verb phrase ellipsis constructions by children at normal and slowed speech rates. Journal of Speech, Language, and Hearing Research, 55, 710–725. Carlson, K. (2001). The effect of parallelism and prosody in the processing of gapping structures. Language and Speech, 44, 1–26. Carlson, K., Dickey, M. W., & Kennedy, C. (2005). Structural economy in the processing and representation of gapping sentences. Syntax, 8, 208–228. Chen, S., & Bargh, J. A. (1999). Consequences of automatic evaluation: Immediate behavior predispositions to approach or avoid the stimulus. Personality and Social Psychology Bulletin, 25, 215–224. Claus, B., & Kelter, S. (2009). Embodied language comprehension: The processing of spatial information during reading and listening. In A.M. Columbus (Ed.), Advances in psychology research, Vol. 59. (pp. 1–44). Hauppauge, NY: Nova Science. de Vega, M., Moreno, V., & Castillo, D. (2013). The comprehension of action-related sentences may cause interference rather than facilitation on matching actions. Psychological Research, 77, 20–30. de Vega, M., & Urrutia, M. (2011). Counterfactual sentences activate embodied meaning: An action–sentence compatibility effect study. Journal of Cognitive Psychology, 23, 962–973. Franklin, N., Tversky, B., & Coon, V. (1992). Switching point of view in spatial mental models. Memory & Cognition, 20, 507–518. Frazier, L., & Clifton, C. J. (2001). Parsing coordinates and ellipsis: Copy α. Syntax, 4, 1–22. Frazier, L., & Clifton, C. J. (2005). The syntax–discourse divide: Processing ellipsis. Syntax, 8, 121–174.

B. Claus / Acta Psychologica 156 (2015) 104–113 Garnham, A. (2001). Mental models and the interpretation of anaphora. Hove. UK: Psychology Press. Gennari, S. P., MacDonald, M. C., Postle, B. R., & Seidenberg, M. S. (2007). Contextdependent interpretation of words: Evidence for interactive neural processes. NeuroImage, 35, 1278–1286. Glenberg, A.M. (1997). What memory is for. Behavioral and Brain Sciences, 20, 1–55. Glenberg, A.M. (2007). Language and action: Creating sensible combinations of ideas. In G. Gaskell (Ed.), Oxford handbook of psycholinguistics (pp. 361–370). Oxford, UK: Oxford University Press. Glenberg, A.M., & Kaschak, M. P. (2002). Grounding language in action. Psychonomic Bulletin & Review, 9, 558–565. Glenberg, A.M., Sato, M., & Cattaneo, L. (2008). Use-induced motor plasticity affects the processing of abstract and concrete language. Current Biology, 18, R290–R291. González, J., Barros-Loscertales, A., Pulvermüller, F., Meseguer, V., Sanjuán, A., Belloch, V., et al. (2006). Reading cinnamon activates olfactory brain regions. NeuroImage, 32, 906–912. Hankamer, J., & Sag, I. A. (1976). Deep and surface anaphora. Linguistic Inquiry, 7, 391–426. Hauk, O., Johnsrude, I., & Pulvermüller, F. (2004). Somatotopic representation of action words in human motor and premotor cortex. Neuron, 41, 301–307. Hoeks, J. C. J., Redeker, G., & Hendriks, P. (2009). Fill the gap! Combining pragmatic and prosodic information to make gapping easy. Journal of Psycholinguistic Research, 38, 221–235. Hofmann, M. (2006). Verarbeitung elliptischer Satzkonstruktionen beim Sprachverstehen. Unpublished PhD thesis. Universität Hamburg. Hommel, B., Müsseler, J., Aschersleben, G., & Prinz, W. (2001). The theory of event coding (TEC): A framework for perception and action planning. Behavioral and Brain Sciences, 24, 849–878. Horton, W. S., & Rapp, D. N. (2003). Out of sight, out of mind: Occlusion and the accessibility of information in narrative comprehension. Psychonomic Bulletin & Review, 10, 104–110. Jackendoff, R. S. (1971). Gapping and related rules. Linguistic Inquiry, 2, 21–35. Johnson, K. (2009). Gapping isn't (VP) ellipsis. Linguistic Inquiry, 40, 289–328. Kaan, E., Overfelt, C., Tromp, D., & Wijnen, F. (2013). Processing gapped verbs. Journal of Psycholinguistic Research, 42, 307–338. Kaan, E., Wijnen, F., & Swaab, T. Y. (2004). Gapping: Electrophysiological evidence for immediate processing of “missing” verbs in sentence comprehension. Brain and Language, 89, 584–592. Kaschak, M. P., Madden, C. J., Therriault, D. J., Yaxley, R. H., Aveyard, M., Blanchard, A. A., et al. (2005). Perception of motion affects language processing. Cognition, 94, B79–B89. Kaschak, M. P., Zwaan, R. A., Aveyard, M., & Yaxley, R. H. (2006). Perception of auditory motion affects language processing. Cognitive Science, 30, 733–744. Kaup, B., Lüdtke, J., & Steiner, I. (2012). Word- vs. sentence-based simulation effects in language comprehension. In B. Stolterfoht, & S. Featherston (Eds.), Empirical approaches to linguistic theory: Studies in meaning and structure (pp. 267–290). Berlin: De Gruyter Mouton. Kilner, J. M., Paulignan, Y., & Blakemore, S. J. (2003). An interference effect of observed biological movement on action. Current Biology, 13, 522–525. Kuno, S. (1976). Gapping: A functional analysis. Linguistic Inquiry, 7, 300–318. Martin, A. E., & McElree, B. (2008). A content-addressable pointer mechanism underlies comprehension of verb-phrase ellipsis. Journal of Memory and Language, 58, 879–906. Martin, A. E., & McElree, B. (2010). Memory operations that support language comprehension: Evidence from verb-phrase ellipsis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35, 1231–1239. Masson, M. E. J., Bub, D. N., & Newton-Taylor, M. (2008). Language-bases access to gestural components of conceptual knowledge. Quarterly Journal of Experimental Psychology, 61, 869–882. Masson, M. E. J., & Loftus, G. R. (2003). Using confidence intervals for graphically based data interpretation. Canadian Journal of Experimental Psychology, 57, 203–220. McKoon, G., & Ratcliff, R. (1986). Inferences about predictable events. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 82–91. Meteyard, L., Bahrami, B., & Vigliocco, G. (2007). Motion detection and motion verbs. Psychological Science, 18, 1007–1013. Meteyard, L., Cuadrado, S. R., Bahrami, B., & Vigliocco, G. (2012). Coming of age: A review of embodiment and the neuroscience of semantics. Cortex, 48, 788–804. Moody, C. L., & Gennari, S. P. (2010). Effects of implied physical effort in sensory-motor and pre-frontal cortex during language comprehension. NeuroImage, 49, 782–793.

113

Nazir, T. A., Boulenger, V., Roy, A., Silber, B., Jeannerod, M., & Paulignan, Y. (2008). Language-induced motor perturbations during the execution of a reaching movement. Quarterly Journal of Experimental Psychology, 61, 933–943. Nicol, J., Swinney, D., Love, T., & Hald, L. (2006). The on-line study of sentence comprehension: An examination of dual task paradigms. Journal of Psycholinguistic Research, 35, 215–231. Pecher, D., Zeelenberg, R., & Barsalou, L. W. (2003). Verifying different-modality properties for concepts produces switching costs. Psychological Science, 14, 119–124. Poirier, J., Wolfinger, K., Spellman, L., & Shapiro, L. P. (2010). The real-time processing of sluiced sentences. Journal of Psycholinguistic Research, 39, 411–427. Pulvermüller, F., & Hauk, O. (2006). Category-specific conceptual processing of color and form in left fronto-temporal cortex. Cerebral Cortex, 16, 1193–1201. Repp, S. (2009). Negation in gapping. Oxford: Oxford University Press. Ross, J. R. (1970). Gapping and the order of constituents. In M. Bierwisch, & K. E. Heidolph (Eds.), Progress in linguistics (pp. 249–259). Den Haag: Mouton. Sag, I. A., & Hankamer, J. (1984). Toward a theory of anaphoric processing. Linguistics and Philosophy, 7, 325–345. Sanford, A. J., Clegg, M., & Majid, A. (1998). The influence of types of character on processing background information in narrative discourse. Memory & Cognition, 26, 1323–1329. Sato, M., & Bergen, B. K. (2013). The case of the missing pronouns: Does mentally simulated perspective play a functional role in the comprehension of person? Cognition, 127, 361–374. Sato, M., Sakai, H., Wu, J., & Bergen, B. K. (2012). Towards a cognitive science of literary style: Perspective-taking in processing omniscient versus objective voice. In N. Miyake, D. Peebles, & R. P. Cooper (Eds.), Proceedings of the 34th Annual Meeting of the Cognitive Science Society (pp. 959–964). Austin, TX: Cognitive Science Society. Shapiro, L. P., & Hestvik, A. (1995). On-line comprehension of VP-ellipsis: Syntactic reconstruction and semantic influence. Journal of Psycholinguistic Research, 24, 517–532. Shapiro, L. P., Hestvik, A., Lesan, L., & Garcis, A.R. (2003). Charting the time course of VP-ellipsis sentence comprehension: Evidence for an initial and independent structural analysis. Journal of Memory and Language, 49, 1–19. Stanley, J., Gowen, E., & Miall, R. C. (2007). Effects of agency on movement interference during observation of a moving dot stimulus. Journal of Experimental Psychology: Human Perception and Performance, 33, 915–926. Steedman, M. (1990). Gapping as constituent coordination. Linguistics and Philosophy, 13, 207–263. Streb, J., Henninghausen, E., & Rösler, F. (2004). Different anaphoric expressions are investigated by event-related brain potentials. Journal of Psycholinguistic Research, 33, 175–201. Taylor, L. J., Lev Ari, S., & Zwaan, R. A. (2008). Inferences about action engage action systems. Brain and Language, 107, 62–67. Taylor, H. A., & Tversky, B. (1992). Descriptions and depictions of environments. Memory & Cognition, 20, 483–496. Taylor, L. J., & Zwaan, R. A. (2008). Motor resonance and linguistic focus. Quarterly Journal of Experimental Psychology, 61, 896–904. Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., et al. (2005). Listening to action-related sentences activates fronto-parietal motor circuits. Journal of Cognitive Neuroscience, 17, 273–281. Wutte, M. G., Glasauer, S., Jahn, K., & Flanagin, V. L. (2012). Moving and being moved: Differences in cerebral activation during recollection of whole-body motion. Behavioural Brain Research, 227, 21–29. Yoshida, M., Dickey, M. W., & Sturt, P. (2013). Predictive processing of syntactic structure: Sluicing and ellipsis in real-time sentence processing. Language and Cognitive Processes, 28(3), 272–302. Zwaan, R. A. (2004). The immersed experiencer: Toward an embodied theory of language comprehension. In B. H. Ross (Ed.), The psychology of learning and motivation, Vol. 43. (pp. 35–62). New York: Academic Press. Zwaan, R. A., Madden, C. J., Yaxley, R. H., & Aveyard, M. E. (2004). Moving words: Dynamic representations in language comprehension. Cognitive Science, 28, 611–619. Zwaan, R. A., & Taylor, L. J. (2006). Seeing, acting, understanding: Motor resonance in language comprehension. Journal of Experimental Psychology: General, 135, 1–11. Zwaan, R. A., Taylor, L. J., & de Boer, M. (2010). Motor resonance as a function of narrative time: Further tests of the linguistic focus hypothesis. Brain and Language, 112, 143–149.

verb distinction in English stress homographs: an ERP study.

Gapping fields on sloping surfaces.

Verb bias and verb-specific competition effects on sentence production.

The theory-of-mind network in support of action verb comprehension: evidence from an fMRI study.

The Effects of Verb Argument Complexity on Verb Production in Persons with Aphasia: Evidence from a Subject-Object-Verb Language.

Failure of distal biceps repair by gapping.

Stem compatibility for cement-in-cement femoral revision: an in vitro study.

Simultaneous Processing of Noun Cue and to-be-Produced Verb in Verb Generation Task: Electromagnetic Evidence.

Interpretation of verb phrase telicity: sensitivity to verb type and determiner type.

An Evaluation of Blood Compatibility of Silver Nanoparticles.

Compatibility groups in Ascobolus immersus - an indication of speciation.

Stimulus-response compatibility with body parts: a study with hands.

Effects of Type of Agreement Violation and Utterance Position on the Auditory Processing of Subject-Verb Agreement: An ERP Study.

Productivity of Noun Slots in Verb Frames.

MRI insert.

Compatibility study between acetylcysteine and some commonly used tablet excipients.

Acquiring and Producing Sentences: Whether Learners Use Verb-Specific or Verb-General Information Depends on Cue Validity.

Cerebellar Cathodal Transcranial Direct Stimulation and Performance on a Verb Generation Task: A Replication Study.

Bedside blood compatibility testing.

Surface acidity and solid-state compatibility of excipients with an acid-sensitive API: case study of atorvastatin calcium.

The retreat from locative overgeneralisation errors: a novel verb grammaticality judgment study.

The sexual compatibility test.

Immunology. Tamarin compatibility.

Compatibility of an Ultraselective Microcatheter and Epirubicin Loaded 300-500-μm DC Bead in Ex Vivo Study.