Give Me Time to Picture That: Effects of Time and Imageability on Effector-Specific Motor Activation in Idiom Processing
Created: Sept. 13, 2022
A great amount of evidence has now been accumulated indicating that many aspects of cognition, including language are closely interconnected with or “grounded” in the sensory-motor modalities. In this respect, an interesting line of research that has not yet been studied in great detail is whether activation of the sensory-motor brain areas can be found when figurative language is processed. Two experiments were conducted to address the following questions: firstly, can evidence of effector-specific motor activation be found in the processing of idioms; secondly, how would such activation unfold in time? We hypothesized that highly imageable idioms are more strongly coded in different modalities and would therefore be processed faster and be more likely to show evidence of activation in the motor cortex. We also hypothesized that the time between the subsequent presentation of two stimuli would affect the response pattern, as complex simulation processes involving the motor modality in the processing of language would take time to unfold and for two conflicting activations to be integrated. Participants saw pairs of idioms presented one by one and had to indicate whether both idioms were familiar. In Experiment 1, participants responded manually by pressing a button. Experiment 2 required an oral response. For both experiments, highly imageable idioms were processed faster than those with low imageability, indicating a processing advantage related to imageability. More importantly, in Experiment 2 there was a three-way interaction: after a stimulus onset asynchrony (SOA) of 3000 ms there was facilitation in highly imageable idioms in the Same Effector condition compared to the Different Effector condition, suggesting that the motor simulation process might need time to unfold.
Language understanding, embodiment, RT., idioms, motor simulation
Traditionally, language has been viewed as an abstract, amo- dal, proposition-based, symbolically operated system with perception and action acting as simple input-output systems (Wilson, 2002). However, there is emerging evidence that sensory-motor representations play a role in the processing of literal language, from individual words to phrases and sentences (see Fischer & Zwaan, 2008, for a review). In particular, a wealth of brain imaging research shows that processing motion-related, effector-specific verbs (e.g., lick, pick, kick) or sentences activates motor and premotor cortices in a somatotopic1 (The motor cortex is organized somatotopically, i. e. specific areas being responsible for specific body parts.) manner (Pulvermüller, Haerle, & Hummel, 2001; Pulvermüller, Hauk, Nikulin, & Ilmoniemi, 2005; Buccino et al., 2005; Tettamanti et al., 2005). Behavioral studies, in turn, have found evidence of activation for a variety of spatial and motor characteristics in response to verbal material (e.g., Glenberg & Kaschak, 2002; Matlock, 2004; Bergen, Lau, Narayan, Stojanovic, & Wheeler 2010). These data are broadly consistent with the claim that cognition is based on action and perception, and is grounded in the human body (Barsalou, 2008). Thus, “embodiment” (that is, the sensory and motor characteristics of the actors body) impacts cognitive processing.
One of the strongest claims of the embodiment theory relates to the processing of abstract language (Jirak, Menz, Buccino, Borghi, & Binkofski, 2010). An obvious extreme example of the abstract language is figurative language, which includes metaphors, idioms, irony and the like. Idioms are of special interest because the combination of meanings of the constituents is usually different from the figurative meaning of an idiom, and the relationship between literal and figurative meanings is considered to be arbitrary (Cacciari et al., 2011). Cacciari et al. (2011) conducted а transcranial magnetic stimulation (TMS) study which confirmed that verb reading engaged the motor system even if the verbs were embedded in fictive motion or metaphoric sentences. However, such activation was not observed for idiomatic sentences. On the other hand, Boulenger, Hauk, and Pulvermüller (2009) obtained fMRI evidence for effector-specific motor activation in the processing of idioms. In contrast, Raposo, Moss, Stamatakis, and Tyler (2009) failed to obtain evidence of motor activation for idioms in another fMRI study. The researchers pointed out that there might be a crucial difference in the explored time window. Boulenger, Roy, Paulignan, Deprez, Jeannerod, and Nazir (2006) also emphasized the importance of timing, suggesting that processes that run in parallel interfere with each other. They observed that if an action verb was presented during a reaching movement, interference occurred (in terms of reaction time), but the movement was facilitated if the word was presented before the movement onset.
Also exploring the temporal dynamics of activation, Borreggine and Kaschak (2006) reproduced the original action-sentence compatibility (ACE) study by Glenberg and Kaschak (2002) but varied the time at which a response cue was given, indicating when participants should initiate a movement. The ACE facilitation effect for directioncompatible actions and sentences was obtained only if the cue was given at the onset of the stimulus sentence, but presenting the cue at 1000 ms after that resulted in a reversal of the pattern, i.e. inhibition. The authors explained the results within the framework of the Theory of Event Coding (TEC) by Hommel, Musseler, Aschersleben, and Prinz (2001). According to this theory, action planning is divided into two stages. First, there is activation of all action relevant features; at this stage, facilitation for processes activating the same motor features can be observed. During the following stage, the features are integrated and bound to a specific action, so inhibition might be observed due to competition for the same resources. This approach is useful and revealing since it differentiates between mere activation and the integration of features, which prompts the question of what other factors, apart from time, might be influencing the integration. Possible candidates would be the nature of the task and the characteristics of the stimuli themselves.
Our previous research (Gradinarova & Janyan, 2011a; Gradinarova & Janyan, 2011b) showed that both imageability (the ease and speed with which a mental image is evoked by a verbal stimulus) and the nature of the task (familiarity and meaning-verification tasks) affected the activation pattern in the processing of idiomatic expressions. Imageability seems to be an important dimension of the representation and processing of motion verbs and idioms. It is closely connected to mental imagery and, therefore, to our sensorimotor experience (Kosslyn, Ganis, & Thompson, 2001; Kosslyn, 2005; see also Lakoff, 1994, for the importance of imagery in discovering the meaning of tropes and idioms). In Gradinarova and Janyan (2011a, 2011b) a semantic interference paradigm was applied as described by Bergen et al. (2010). Action picture-verb pairs were presented. Participants were asked to indicate whether the picture and the verb stood for the same action or not. The target pairs were those in which both the picture and the verb referred to a specific action performed either with the same effector (e.g., foot in the pair kick-run) or a different effector (e.g., foot and hand in the pair jump-write). Analysis showed that participants were slower in saying “No” in the same-effector condition than in the different-effector condition. Bergen et al. (2010) accounted for this interference effect in terms of competing activation of the same motor circuits.
The same logic was used in Gradinarova and Janyan (2011a, 2011b) but verbs and idioms were used as stimuli instead. Pictures were not included because automatic motor activation in linguistic processing should also be elicited when only linguistic stimuli are presented, and such a presentation method would lead to greater simplicity of the experiment and exclude an additional modality to be studied. Secondly, finding images to depict phrases with idiomatic, non-literal meanings would be difficult and would unnecessarily strengthen the subjectivity factor as different participants might imagine a situation described by an idiom in different ways. To verify the strong embodiment claim, Gradinarova and Janyan (2011b) used idiom pairs as stimuli, with the first idiom always containing a verb referring to a hand-action and the second idiom containing either a hand-action verb or a verb referring to a different-effector action. Imageability was hypothesized to be a factor of importance. Following the logic of Paivios (1991) Dual Coding Theory, which states that language is represented in two separate but strongly interconnected systems - logogens (verbal) and imagens (non-verbal, sensory) - it was supposed that highly imageable idioms would be strongly represented in the sensory-motor system (see also Boulenger et al., 2009) and would have a processing advantage as well as being more likely to elicit stronger motor activation effects. Participants had two different tasks in two separate experiments: meaning verification (indicating whether two idioms have the same meaning) and familiarity verification (indicating whether both idioms were familiar to them). In both experiments, highly imageable idioms elicited significantly faster responses (supporting the hypothesized processing advantage) as well as a semantic interference effect. Low-imageability idioms were influenced by the task and gave an opposite, facilitation effect. That is, the same-effector idioms were processed faster than different-effector idioms in the familiarity verification task, but no significant effect was found in the meaning verification task. Importantly, experiments that reversed the idiom presentation order so that a response was always given to a phrase involving an action with the hand effector (Markova, 2011) failed to reproduce the effects. Unpublished data from recent research conducted by Gradinarova and Janyan (2011) suggests that highly imageable idiomatic phrases presented individually actually interfere with the hand movement required for the response (pressing a button).
As in our 2011b study, the experiments presented here implemented the semantic interference paradigm and varied imageability as a factor possibly influencing motor activation for idiomatic language. Additionally, the methodological corrections from Markova (2011) were incorporated. A familiarity verification task was implemented to test the previously obtained facilitation effects (in low-imageability idioms) versus inhibition effect (in high-imageability idioms) and whether these effects might have resulted from methodological faults. An SOA manipulation was included so as to test for temporal effects. A different task was also used in Experiment 2 in order to preclude any effect resulting from the manual response action interacting with the idiom processing. Our main focus was on highly imageable idioms. We hypothesized that: 1) highly imageable idioms would have a processing advantage; 2) stimuli with higher imageability would elicit stronger motor activation, i.e. there would be a difference in reaction times (RTs) for same-effector versus different- effector pairs of highly imageable idioms; 3) this difference might be influenced by the temporal dynamics of activation, though the exact pattern of the activation was hard to predict given the scarce amount of previous research.
Participants: 25 native Bulgarian speakers (18 females) participated in the experiment (mean age=24.6, SD=4.6).
Stimuli and Design: A 2 (Effector Matching: Same vs. Different) x 2 (Imageability: High vs. Low) x 2 (SOA: 2000 ms vs. 3000 ms) factorial design was used. The stimuli were the same as in Experiment 2 in Gradinarova and Janyan (2011b) but the order of presentation within an idiom pair was reversed for methodological reasons (so that the response would always follow an idiom containing a HAND-action verb). Target stimuli consisted of 16 pairs of high-imageability and 16 pairs of low-imageability idioms, controlled for their familiarity, transparency and length (see Table 1). All idioms had a similar syntactic structure, consisting of a verb that referred to an action performed with a specific effector (HAND, FOOT, MOUTH) and a noun and/or prepositional phrase. The second idiom in a pair always contained a HAND-action verb. It was presented after a phrase with a verb referring either to another HAND-action or to an action with a different effector (see Figure 1). SOA was varied so that after a 1500 ms presentation of the first idiom, the interstimulus interval was either 500 ms or 1500 ms. 18 more pairs of Bulgarian idioms and 50 pairs containing idioms translated from another language were used as fillers.
Procedure: Participants were asked to discriminate between familiar and unfamiliar idioms. Two idioms appeared one after another on the screen. Participants were instructed to give a “YES” response if BOTH of these expressions were familiar to them, and a “NO” response if ANY (either or both) of the idioms were unfamiliar to them. The task was implemented in order to engage the participants in carefully reading and trying to understand the stimuli. Motor activation was expected to occur automatically with the processing of the linguistic stimuli and to a different degree depending on the characteristics of these stimuli.2 (Unpublished research by Gradinarova and Janyan (2011) suggests that the effects elicited when implementing the familiarity judgment task cannot be explained merely by lexical retrieval processes. The processing of the HAND-action verb in the stimuli seemed to interact with the hand-action (pressing a button) required for the response, indicating involvement of motor activation in language processing.) All fillers required a “NO” response and all “NO” responses to target (familiar) stimuli were considered erroneous (erroneous responses included errors, in which participants accidentally indicated that they did not know an idiom that was actually familiar to them, as well as cases in which the target stimuli were unfamiliar to the participants and were therefore not analyzed, since the focus of the present study is on familiar idioms. Only “YES” responses were analyzed. Four lists were devised in order to counterbalance the Effector-Matching and SOA conditions across idioms. Each participant saw each target idiom only once, as the idioms were randomly assigned to one of the four lists. The stimuli presentation was pseudorandomized, so that the same condition did not appear in more than three consecutive trials.
Participants were given written and oral instructions, after which they went through a short practice session. Each trial started with a fixation cross (“+”) for 1000 ms, followed by an idiom, which stayed on the screen for 1500 ms, followed by a blank screen for either 1500 ms or 500 ms (depending on the SOA condition), and then by another idiom, which stayed on the screen for 4000 ms or until a response was given. The inter-trial interval was 1500 ms. E-Prime software (Schneider, Eschman, & Zuccolotto, 2002) was used to present the stimuli and record the response accuracy and RT.
Prior to the analysis3 (Although there was not an equal amount of subject data per list we report data of all 25 participants. Data of 24 participants were analyzed and the results were of no difference from the 25 subject data. Due to space restriction we report only one analysis.) , erroneous responses (7.1%) and response times lying more than ± 2.5 standard deviations from the mean per condition (2.1%) were excluded.
A 2 (Effector Matching: Same vs. Different) x 2 (Imageability: High vs. Low) x 2 (SOA: 3000 ms vs. 2000 ms) repeated measures ANOVA was performed for subject and item means. Table 2 presents the means and standard deviations for each condition.
The analysis revealed a significant main effect of Imageability in subject and item means (Fs(l,24)=16.59, p=.00, r|p=.41; F.(l,26)=5.21, p=.O3, r]p2=.17). Response times to high-imageability idioms were faster than for low-imageability idioms (see Table 2). This main effect confirms the results of Gradinarova and Janyan (2011a, 2011b), who reported faster response times for high-imageability idioms than for low-imageability idioms. In addition, a marginal effect of SOA in both subject and item means was revealed (Fs(l,24)=3.62, p=.O7, r|p2=.13; F.(l,26)=4.13, p=.O5, r|p2=.14), suggesting that subjects were slower in the 3000 ms SOA condition than in the 2000 ms SOA condition. A marginal interaction between Effector Matching and SOA was also found (Fs(l,24)=3.30, p=.O8, r|p2=.12; F.(l,26)=4.03, p=.O6, r|p2=.13). The Fischer LSD post-hoc test showed that at SOA 3000 ms, same-effector idioms were processed significantly slower than at SOA 2000 ms (p=.01).
As a whole, Experiment 1 did not produce any convincing evidence for effector-specific motor activation in both types of idioms, although the marginal interaction between Effector Matching and SOA did suggested an inhibition effect at a later time window for phrases sharing the same effector. However, a possible interaction between the manual response movement and the processing of the idioms might have masked any effects. Therefore, Experiment 2 was designed to preclude any interaction of that kind, by changing the response mode from manual to oral.
Participants: 20 native Bulgarian speakers (11 females) participated in the experiment (mean age=8.1, SD=7.1).
Stimuli and Design: The stimuli and design were identical to Experiment 1.
Procedure: The same as in Experiment 1, the only difference being that responses were oral (“Yes’7”No”). A serial response button box recorded voice onset RT through a microphone.
Prior to the analysis, erroneous responses (7.7%) and response times lying more than ± 2.5 standard deviations from the mean per condition (2.3%) were excluded. A repeated measures ANOVA was performed for subject and item means (see Table 3 for means and SDs for each condition).
As with Experiment 1, the analysis revealed a significant main effect of Imageability in subject means but not in item means (Fs(l,19)=20.66, p=.00, r) 2=.52; F.( 1,30) =2.93, p=.l). The main effect of SOA was also significant both in subject means and in item means (F(l,19)=6.17, p=.O3, f]p2=.25; F.(l,30)=17.76, p=.00, r) 2=.37). More importantly, two two-way interactions were obtained between Effector Matching and Imageability (F(l,19)=7.83, p=.O2, np2=.29; F.(l,30)=4.58, p=.O4, r) 2=.13) and between SOA and Imageability in item means only (Fs(l,19)=1.42, p=.2; F.(l,30)=4.78, p=.O4, r) 2=.14). Most crucially, there was a significant three-way interaction (F (1,19)=8.18, p=.O2, t]p2=.3O; E(l,30)=5.17, p=.O3, r|p2=.15), plotted in Figure 2. Fischers LSD post-hoc test revealed a significant difference between the Effector Matching conditions for high-imageability idioms in the 3000 ms SOA condition (p<.05) both in subject and in item means, suggesting a facilitation effect for highly imageable idioms sharing a same-effector verb for 3000 ms only.
Thus, the results from Experiment 2 suggested that: (i) participants were slower after a longer SOA; (ii) Imageability and Effector matching interacted, indicating that highly imageable idioms might trigger a stronger motor simulation; (iii) SOA interacted with the other factors, supporting the hypothesis that the temporal unfolding of activation is important, and more specifically, that at a later stage and for highly imageable idioms, features activated by the first expression in a pair facilitate the processing of the second idiom. Low imageability idioms remained indifferent to the experimental manipulations. Since our focus was on highly imageable idioms as bearing more potential for revealing motor activation we exclude the null results of low imageability idiom from further discussion.
The comparatively new concept of embodied cognition provides a theoretical framework for the interaction of high- and low-level processes. It links personal experience, sensorimotor processes and cognitive processes such as language comprehension. The theory claims that language processing recruits the same sensorimotor areas needed for action execution or interaction with objects to which the words refer One of the strongest claims of embodiment cognition refers to the embodiment of abstract language. Experimental data on the link between motor representation and abstract, idiomatic language are scarce and inconclusive.
The current experiments aimed to explore a possible motor activation in the processing of idioms by using the semantic interference paradigm (Bergen et al., 2010). Another aim of the study was to trace the time needed for capturing the motor activation. Both experiments used the same paradigm and the same target stimuli. Two idioms were presented to the participants that contained verbs for actions performed with a particular effector (hand, foot, mouth). The first idiom either shared the same effector as the second one (hand) or did not. Based on previous rating studies the stimuli were evenly divided into pairs of idioms with either high or low imageability. Two SOAs were used: 2000 ms or 3000 ms. We hypothesized that there would be a significant difference between the RTs if idioms activated the same motor circuit in comparison with the motor circuits responsible for different effectors. Based on our previous research (Gradinarova & Janyan, 2011a; 2011b), a processing advantage as well as an interference effect were expected to appear for highly imageable idioms. However, we had no specific prediction concerning motor activation timing due to the fact that the variety of previous research mentioned in the introduction has not yet led to the formulation of a systematic and clear principle describing the influence of time. We did hypothesize that the activation and integration processes would be influenced by the time given for their unfolding.
In keeping with Paivio’s (1991) Dual Coding Theory and in accordance with various imagery data (Kosslyn et al. 2001; Kosslyn, 2005), we found a processing advantage for highly imageable idioms which are presumably well-represented not only in a verbal system but also in the sensory-motor system (Boulenger et al., 2009). In Experiment 1, any effects of the activation of overlapping motor circuitry might have been masked by the hand motor task. In Gradinarova and Janyan (201 lb) an inhibition effect was observed for highly imageable idioms. The response was made by pressing a button following either an idiom containing a HAND-action verb or an idiom containing a FOOT-/MOUTH-action verb and thus the inhibition might have been due to competition between the motor activation triggered by the HAND-action idiom and the HAND-action response. Similarly, the current study’s Experiment 1 revealed a tendency toward inhibition when same-effector idioms were presented with an SOA of 3000 ms. Experiment 2 excluded the possibility of inhibition by introducing an oral response. Facilitation was then observed at a later time window. The facilitation effect (RT difference between the same and different effector conditions) was valid for highly imageable idioms only. It might be that, given enough time to unfold, the motor activation from the first phrase facilitated the processing of the second. This might be explained in terms of better integration given a greater time difference between the stimuli, a hypothesis that would be in keeping with the TEC (Hommel et al., 2001). In terms of the TEC, the observed facilitation may reflect activation of all actionrelevant features during the processing of both idioms. The TEC would predict that the facilitation effect could be reversed into inhibition at an even later stage, when the features activated by the first idiom would be integrated and the features activated by the second idiom would enter into competition within the same circuitry. The lack of an earlier effect could be a result of a slower activation of motor features in the idiomatic phrases. However, when a HAND-response was required, in addition to the two idiomatic phrases, which already had integrated features, response would interfere strongly with the two previous activations of the hand circuitry and would result in inhibition. This could account for the presence of inhibition rather than facilitation with the manual response. An additional possibility is that it is generally more difficult for motor activations accompanying actual manual responses to be integrated with motor activations from language processing. Thus, the present study suggests that researchers should be careful when implementing a motor task in experiments in the grounded cognition paradigm.
An alternative explanation for the facilitation effect could be that it was due to pure lexical associations between words that share the same effector. That is, the verb stir embedded in an idiom would spread the activation over the lexico-semantic system and activate all lexically associated words extend being among them. Then, given time, lexical associations might have been triggered resulting in faster recognition of extend within an idiom and, hence, overall processing facilitation compared to the different
effector condition. There are observations, however, that speak against such an interpretation. First, a lexical association effect would be obtained in low imageability idiom processing, too; this is not the case. Second, it is not clear why a lexical association effect would disappear if a manual response was required. An interaction of motor activations could, however, account for this sort of result.
Further research is needed to resolve the questions raised by the current study. The differences between the results of the two experiments presented here suggest that the processes of motor activation and motor integration might be complicated and dependent on a variety of factors, such as time, word properties such as imageability, and the nature of the task implemented. Special attention should also be devoted to the methodology used in experiments in this field, as demonstrated by the consequences of implementing an interfering motor task, as well as differences in the order of stimuli presentation. To conclude, besides methodological issues, the studies provided some initial evidence for the timeline of the grounding of idiomatic language and of sensorimotor involvement in abstract idiomatic language processing that can be captured by response times.
Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617-645. doi: 10.1146/annurev. psych.59.103006.093639
Bergen, B., Lau, T. T. C., Narayan, S., Stojanovic, D., & Wheeler, K. (2010). Body part representations in verbal semantics. Memory & cognition, 38(7), 969-981. doi: 10.3758/MC.38.7.969
Borreggine, K. L., & Kaschak, M. P. (2006). The action-sentence compatibility effect: It’s all in the timing. Cognitive Science, 30, 1097-1112. doi: 10.1207/sl5516709cog0000 91
Boulenger, V., Hauk, O., Pulvermüller, F. (2009). Grasping ideas with the motor system: semantic somatotopy in idiom comprehension. Cerebral Cortex, 19(8), 1905-1914. doi: 10.1093/cercor/bhn217
Boulenger, V., Roy, A. C., Paulignan, Y., Deprez, V., Jeannerod, M., & Nazir, T. Ä. (2006). Cross-talk between language processes and overt motor behavior in the first 200 ms of processing. Journal of Cognitive Neuroscience, 18(10), 1607-1615. doi: 10.1162/jocn.2006.18.10.1607
Buccino, G., Riggio, L., Meili, G., Binkofski, F., Gallese, V., & Riz- zolatti, G. (2005). Listening to action-related sentences modulates the activity of the motor system: a combined TMS and behavioral study. Cognitive Brain Research, 24(3), 355-363. doi: 10.1016/j.cogbrainres.2005.02.020
Cacciari, C., Bolognini, N., Senna, I., Pellicciari, M. C., Miniussi, C., Papagno, C. (2011). Literal, Active and metaphorical motion sentences preserve the motion component of the verb. A TMS study. Brain and Language, 119(3), 149-157. doi: 10.1016/i.bandl.2011,05.004
Fischer, M. H. & Zwaan, R. A. (2008). Embodied language: a review of the role of the motor system in language comprehension. Quarterly Journal of Experimental Psychology, 61, 825-850. doi: 10.1080/17470210701623605
Glenberg, A. M. & Kaschak, M. P. (2002). Grounding language in action. Psychonomic Bulletin d* Review, 9(3), 558-565. doi: 10.3758/BF03196313
Gradinarova, G. & Janyan, A. (2011a). Effector-specific motor simulation in idiom processing. In B. Kokinov, A. Karmil- off-Smith, N.J. Nersessian (Eds.), European Perspectives on Cognitive Science. Sofia: New Bulgarian University Press.
Gradinarova, G. & Janyan, A. (2011b). Motor simulation and verbal association in idiom-idiom verification: Effects of imageability. In L. Carlson, C. Hoelscher, & T.E Shipley (Eds.), Proceedings of the 33rd Annual Conference of the Cognitive Science Society (pp. 3052-3057). Austin, TX: Cognitive Science Society.
Hommel, B., Musseier, J., Aschersleben, G., & Prinz, W. (2001). The theory of event coding (TEC): A framework for perception and action planning. Behavioral and Brain Sciences, 24, 849-937. doi: 10.1017/S0140525X01000103
Jirak, D., Menz, M. M., Buccino, G., Borghi, A. M., & Binkofski, F. (2010). Grasping language - A short story on embodiment. Consciousness and cognition, 19(3), 711-720. doi: 10.1016/j.concog.2010.06.020
Kosslyn, S. M. (2005). Mental images and the brain. Cognitive Neuropsychology, 22(3/4), 333-347. doi: 10.1080/ 02643290442000130
Kosslyn, S. M., Ganis, G., & Thompson, W. L. (2001) Neural foundations of imagery. Nature Reviews Neuroscience, 2, 635- 642. doi: 10.1038/35090055
Lakoff, G. (1994). What is metaphor? In J. Bardnen & K. Holyoak (Eds.), Advances in Connectionist and Neural Computational Theory. Analogy, Metaphor, and Reminding (pp. 203-258). Vol.3. Norwood, NJ: Ablex.
Markova, V. (2011). Motor simulations in idiom processing (in Bulgarian, Unpublished manuscript) [Bachelor thesis] Sofia: NBU.
Matlock, T. (2004). Fictive motion as cognitive simulation. Memory & Cognition, 32, 1389-1400. doi: 10.3758/BF03206329
Paivio, A. (1991). Dual coding theory: Retrospect and current status. Canadian Journal of Psychology, 45(3), 255-287. doi: 10.1037/h0084295
Pulvermüller, E, Hauk, O., Nikulin, V. V., & Ilmoniemi, R. J. (2005). Functional links between motor and language systems. European Journal of Neuroscience, 21(3), 793-797. doi: 10.1111/i. 1460-9568.2005.03900.X
Pulvermüller, E, Häerle, M. & F. Hummel (2001). Walking or talking?: Behavioral and neurophysiological correlates of action verb processing. Brain and Language, 78(2), 143-168. doi: 10.1006/brln.2000.2390
Raposo, A., Moss, H. E., Stamatakis, E. A. & Tyler, L. K. (2009). Modulation of motor and premotor cortices by actions, action words and action sentences. Neuropsychologia, 47, 388-396. doi: 10.1016/j.neuropsychologia.2008.09.017
Schneider, W, Eschman, A., & Zuccolotto, A. (2002). E-Prime: Users guide. Psychology Software Incorporated.
Scorolli, C., Borghi, A. M. (2007). Sentence comprehension and action: Effector specific modulation of the motor system. Brain Research, 1130, 119-124. doi: 10.1016/j. brainres.2006.10.033
Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., Fazio, E, Rizzolatti, G., Cappa, S. E, Perani, D. (2005). Listening to action-related sentences activates fronto-parietal motor circuits. Journal of Cognitive Neuroscience, 17, 273-281. doi: 10.1162/0898929053124965
Wilson, M. (2002). Six views of embodied cognition. Psychonomic Bulletin & Review, 9(4), 625-636. doi: 10.3758/BF03196322