The effect of executive control ability on the comprehension of second language metaphor

Abstract

Aims:

The present study examined the effect of language-related executive control ability on second language (L2) metaphor comprehension in L2 learners.

Design:

All participants were Chinese–English bilinguals. The Stroop task was used to measure language-related executive control ability. Three types of sentences were used as stimulus materials, including familiar metaphoric sentences, unfamiliar metaphoric sentences and literal sentences. Participants were asked to determine whether the sentence presented was metaphoric or not.

Data and analysis:

Both response latencies and accuracy scores were obtained. Linear mixed effect model was used for statistical analysis.

Findings:

The effect of executive control ability on L2 metaphor comprehension is modulated by the familiarity of the metaphor. Specifically, for familiar metaphor sentences, the response time of participants with higher executive control ability was significantly faster than those with lower executive control ability. However, for the unfamiliar metaphor sentences and literal sentences, the effect of executive control ability was not significant.

Originality:

This study directly explored the impact of language-related executive control on metaphor comprehension in L2 learners.

Significance/Implications:

Our results found that the Predication Model is also applicable to L2 learners.

Keywords

Executive control second language metaphor comprehension familiarity Predication Model

Introduction

The use of metaphor is very common in our daily life, enriching our language expression and making it more vivid. Native speakers use metaphors constantly, and so do second language (L2) learners, who are also exposed to metaphoric expressions during their language acquisition and application. Because of the cognitive and cultural attributes of metaphors, their usage in L2 teaching is assumed to promote L2 learning. Therefore, L2 metaphor comprehension is important to L2 learners. Existing studies on native speakers found that individual differences exist for metaphor comprehension (Chiappe & Chiappe, 2007; Kazmerski, Blasko, & Dessalegn, 2003; Mashal, 2013; Pierce, MacLaren, & Chiappe, 2010). Are there any individual differences for L2 metaphor comprehension as well? This study would focus on the impact of L2 learners’ executive control ability on their metaphor comprehension.

In studies on native speakers, the process of metaphor comprehension has attracted the attention of researchers. The Predication Model (Kintsch, 2000, 2001) suggested that metaphor can be properly interpreted by a computational process, and it uses two components: the latent semantic analysis (LSA) component and the construction-integration (CI) component. LSA represents the meaning of words in a multidimensional semantic space, with vectors to represent their positions, and the cosine of their vectors to represent the semantic relations. The CI uses these LSA vectors to construct the interpretation of statements. To comprehend a metaphor, for example, ‘My lawyer (Argument) is a shark (Predicate),’ the first thing for CI is to activate the m closest properties in the semantic neighborhood of “shark (predicate),” and then to modify the vector of “shark” by a spreading activation process in an inhibitory network to yield a contextually appropriate meaning. The network contains Argument (A), Predicate (P) and the m nearest semantic neighbors of P. Properties which are related to both A and P are enhanced (e.g. predatory, aggressive, and tenacious), but properties which are not related to A are inhibited (e.g. fast swimmer, fish, sharp teeth, and leathery skin). In these properties, the K properties with the greatest activation characterize the meaning, resulting in a proper interpretation. Following this line of reasoning, we can speculate that the ability to inhibit irrelevant information will directly affect metaphor comprehension.

Second language metaphor comprehension may involve not only the inhibition of irrelevant information, but also the inhibition of non-target language information (Mercier, Pivneva, & Titone, 2014). Therefore, the executive control ability may be more important in L2 metaphor comprehension. Then one question arises: whether L2 metaphor comprehension is affected by individuals’ executive control ability or not. In the following section, we explained first the factors that affect first language (L1) metaphor comprehension, which paves way for L2 metaphor comprehension.

Individual differences and L1 metaphor comprehension

Studies on native speakers have found that the central executive control component of working memory (WM) plays an important role in the comprehension of metaphor (Chiappe & Chiappe, 2007; Kazmerski et al., 2003; Mashal, 2013; Olkoniemi, Ranta, & Kaakinen, 2016; Pierce et al., 2010).

Chiappe and Chiappe (2007) explored the effects of WM and specifically analyzed which component of WM affected metaphor comprehension. In this study, half of the sentences used were difficult (i.e. metaphors with low-constraint topics and high-ambiguity vehicles) and half were easy (i.e. metaphors with low ambiguity vehicles and high-constraint topics). WM was measured using the vocabulary auditory span task and inhibitory control by a Stroop task. All participants were asked to explain the metaphoric sentences presented on the screen, while pressing a key to record the time. The results showed that individuals with higher WM span had better inhibitory control, and they interpreted metaphors much faster and with much higher quality. However, this relationship between WM span and the interpretation speed was reduced by about 1/3 when taking the inhibitory response time of Stroop task as a covariate. This indicates that executive control plays an important role in-between. Specifically, the difference in the time of metaphor interpretation could be partly attributed to the difference in executive control.

Mashal (2013) explored the relationship between metaphor comprehension and WM using materials with different familiarity (Experiment 1). Both forward and backward digital WM span were measured. The backward digital span reflects central executive inhibitory control (St Clair-Thompson & Gathercole, 2006). Results showed that the comprehensions of both familiar and unfamiliar metaphors were positively correlated with the backward digit span, but not with the forward digit span, indicating that the central executive system is involved in both types of metaphor comprehension. Pierce et al. (2010) examined the role of WM in the metaphoric interference effect which indicates automatic activation of metaphoric meaning (MIE, i.e. participants spending longer time to judge whether metaphors were literally false compared to scrambled sentences). The participants were asked to determine whether the metaphoric sentences presented were literally correct. Their WM spans were measured using the word span task (WSPAN). In the WSPAN task, one needs to manage proactive interference, that is, to inhibit the previous words so that they do not interfere with the memory of the current set. Specifically, the WSPAN measures proactive interference, which can predict the ability to suppress unwanted thoughts (Verwoerd, Wessel, & de Jong, 2009). The results showed that the MIE of individuals with high WSPAN was significantly lower than that of individuals with low WSPAN, indicating that despite an automatic activation of metaphoric meaning, the early processing of metaphors is influenced by executive mechanisms.

Carriedo et al. (2016) explored the development of metaphor comprehension and its relationship with verbal reasoning and executive function. This study examined the execution function from the perspective of updating, inhibition ability, and cognitive flexibility. The updating ability and cognitive inhibition were measured by working memory task. The response-distractor inhibition was measured by the go/no-go task and the flanker task, and the cognitive flexibility was measured by a modified flanker task. Metaphor comprehension task required participants to explain metaphors. The correlational analysis and the variance communality analysis found that the effect of executive function on metaphor comprehension was obvious at age 15, and it was limited to information updating in working memory and cognitive inhibition. At the same time, the executive function only had obvious promotion effect on the participants with low metaphoric comprehension efficiency.

In summary, previous studies on L1 metaphor have demonstrated the effect of WM, especially the central executive control component, on metaphor comprehension. Up until now, few studies have directly manipulated the executive control ability of individuals to explore its influence on metaphor comprehension. In this regard, Columbus et al. (2015) investigated the influence of domain-general executive control on metaphor comprehension using eye movement measures. They asked the participants to read the metaphorical sentences with different familiarity and context. Their domain-general executive control ability was measured by the AX-CPT task in which letter stimuli that do not depend on language processing is used and the subjects are required to press “yes” button when an ‘X’ immediately follows an ‘A’, and to press “no” button for all other trials. Results showed that individuals with high executive control had longer verb gaze duration and longer metaphor verbs reading times when a prior context forced a particular interpretation (metaphorical or literal), and the total metaphor reading times were shorter when there was a prior context. While the participants with low executive control ability were not influenced by the prior context, they were more likely to regress back into the prior context than those with high executive control, and their metaphor rereading time was longer, indicating that they needed more effort to comprehend the sentence. This finding suggests that individuals with high domain-general executive control ability can integrate the meaning into the previous context more effectively, thus can promote metaphor comprehension. However, it should be noted that the AX-CPT task used in this study is more complex and mainly involves domain-general executive control, such as proactive control and reactive control. Thus, this task could not measure the exact inhibitory process during metaphor comprehension which mainly involves the inhibition of irrelevant information about literal or metaphoric meaning. Language-related executive control may involve in the process of metaphor comprehension. So far, the relationship between domain-general executive control and language-related executive control is still controversial. Some studies have shown that the measures of linguistic and non-linguistic executive control are lack of connection (Branzi, Calabria, Boscarino, & Costa, 2016). The neuroimaging data also revealed incomplete overlap between the two types of executive control (e.g. Blanco-Elorrieta & Pylkkänen, 2016; Branzi et al., 2016; Magezi, Khateb, Mouthon, Spierer, & Annoni, 2012), which means that they do not share exactly the same mechanism. Language-related executive control (as measured by the Stroop task) is related to the suppression of the meaning of the word, and it contains both linguistic factors and executive control. In this sense, it might share the same mechanism as the inhibition of metaphor comprehension. Therefore, we aimed to narrow down to the language-related executive control, and examine its impact on metaphor comprehension in the present study.

Familiarity and metaphor comprehension

Another important factor influencing the comprehension of metaphor is familiarity. Some metaphors are familiar, and others are relatively unfamiliar, so their comprehension process may not be the same. Unfamiliar metaphors may be comprehended by immediate contrast, whereas familiar metaphors are comprehended by extracting classification knowledge from memory (Bowdle & Gentner, 2005). Many studies have shown that familiar metaphor processing is much faster and more accurate than unfamiliar metaphors (Goldstein, Arzouan, & Faust, 2012; Lai & Curran, 2013; Mashal, 2013; Mashal & Faust, 2009). This is consistent with the findings that familiar metaphors are faster primed (Blasko & Connine, 1993) and judged in semantic judgment tasks (Goldstein et al., 2012; Mashal & Faust, 2009).

Studies have also shown that cognitive efforts to comprehend metaphors with different familiarity are not the same (Columbus et al., 2015; Mashal, 2013). Mashal (2013) investigated the relationship between WM and familiar and unfamiliar metaphors using a free recall task (Experiment 3). She found that unfamiliar metaphors are more difficult to remember than familiar ones. Moreover, she assessed the type of errors committed by participants in metaphor recognition tasks and examined the relationship between these errors and working memory (Experiment 2). Phonological errors were used to reflect retrieval from the phonological loop. Semantic errors were used to reflect retrieval of semantic cues. The results showed that unfamiliar metaphor recognition induced more semantic errors, whereas familiar metaphors induced equal rates of phonological errors and semantic errors. These results indicated that unfamiliar metaphor recognition may involve a more elaborative process about selecting related semantic features and suppressing irrelevant semantic information. Familiar metaphors may be remembered through a more superficial process of rehearsing within the phonological loop. Columbus et al. (2015) found that participants’ gaze duration for less familiar metaphoric verbs is longer than verbs with literal meaning, and this difference decreases with increasing metaphor familiarity. However, there were no significant effects for high familiar metaphor verbs. These results indicate that individuals need to expend greater effort to comprehend unfamiliar metaphors.

Taken together, the effect of executive control ability on metaphors with different familiarity may also be different. This assumption has been supported by some previous studies. Mashal (2013) found that the central executive system is involved in both types of metaphor comprehension (Experiment 1). She also found the recognition scores of unfamiliar metaphors are positively correlated with backward digit spans, and the semantic errors are negatively correlated with backward digit spans, indicating that the executive control component is indeed involved in unfamiliar metaphor recognition (Experiment 2). Brain imaging studies have also supported this view. Mashal, Faust, Hendler, & Jung-Beeman (2007) found that unfamiliar two-word metaphors (e.g. sweet sleep) led to greater neural activation in frontal brain regions (the brain regions which are associated with the executive function) compared to familiar metaphors and literal phrases. This finding is not only true for the normal population, but also for individuals with executive control deficits (e.g. people with schizophrenia). Mashal, Vishne, Laor, & Titone (2013) have found that the left middle frontal gyrus (left MFG) of individuals who have deficits in executive control (schizophrenia) has stronger activation in processing unfamiliar metaphors than familiar metaphors.

In summary, we have known that the effect of the executive control component of WM on metaphor comprehension varies with metaphor familiarity. Specifically, the comprehension of unfamiliar metaphors requires a higher level of executive control ability. This finding can be explained by the Predication Model. That is, the comprehension of unfamiliar metaphors requires activating more semantic neighbors than familiar metaphors, so that individuals need to inhibit more unrelated attribution.

The comprehension of L2 metaphors

There are some differences in metaphor comprehension between L2 learners and native speakers. First, native speakers comprehend figurative language effortlessly, while L2 learners often have great difficulty in doing so (e.g. Kecskes, 2006; Littlemore, Chen, Koester, & Barnden, 2011). Littlemore et al. (2011) found that metaphors account for 40% of the more difficult items that were evaluated by participants whose native languages are not English. In addition, in the misinterpreted metaphors, only 4% of the cases were rated difficult. So, it is obvious that L2 learners may not only perceive L2 metaphors as more difficult, but also underestimate this difficulty. The reason might be that the figurative meaning of an utterance is grounded in the socio-cultural experience of the native speakers, from which a particular highly accessible interpretation emerges. Due to the lack of (or limited) experience with the language and the culture, what is accessible for L1 speakers will not necessarily be accessible for L2 speakers (Kecskes, 2006). Secondly, the processing advantage and dominant hemisphere differ between L1 and L2 metaphors. In processing familiar metaphors, there is a left hemisphere advantage in L1 with a fine coding approach, but a right hemisphere advantage in L2 with a coarse semantic coding approach (Mashal, Borodkin, Maliniak, & Faust, 2015). Finally, metaphor saliency differs between native speakers and L2 learners. Native speakers tend to take a holistic view toward figurative language and consider their figurative meaning salient, while L2 learners use an analytic approach toward figurative language and usually take their literal meaning as salient (Kecskes, 2006).

To answer the question what kind of factors may affect the process of metaphor comprehension for L2 learners, the existing research mainly focuses on individual differences and language-related factors such as the degree of salience, L1–L2 similarity and so on. For example, Kecskes (2006) discussed the impact of the salience of linguistic materials on L2 metaphor comprehension. Meanwhile, many recent studies have also focused on the effects of L1–L2 similarity on L2 metaphoric comprehension, such as the similarity of lexical representations and conceptual representations of metaphors between L1 and L2, or the similarity of physical experience of metaphor in the two languages. Some researchers try to explore the difficulties that L2 learners may encounter through manipulating the similarity of the two languages, and predict that the higher the similarity, the less difficult to comprehend (Charteris-Black, 2002; Ferreira, 2008; Safarnejad, Ho-Abdullah, & Awal, 2014; Türker, 2016). However, so far little attention has been paid to the influence of metaphor familiarity on L2 metaphor comprehension.

Studies on the effect of individual differences on L2 metaphor comprehension mainly focus on L2 proficiency, vocabulary size, writing achievement, and cognitive style. For example, Cheung (2013) studied the L2 metaphoric comprehension in Cantonese–English bilinguals, and found that both individual variables and experimental settings influenced metaphor interpretation. Individual variables included L2 proficiency, natural age, different perceptions of general experience, and individual L1–L2 vocabularies. Participants with higher English proficiency performed better in metaphor comprehension, and older children performed better than younger children. Likewise, L2 learners performed better when their L1–L2 vocabulary was larger and their experience was more abundant. Jin’s (2011) study also found that L2 proficiency influences how learners are affected by their first languages: learners with lower proficiency were more affected. Sandgren (2014) found that L2 metaphor comprehension mainly depends on the L2 learners’ knowledge, such as world knowledge, professional knowledge, topic-related knowledge, and has nothing to do with their L2 proficiency. Hoàng (2015) investigated the comprehension and production of metaphor in L2 learners’ articles, focusing primarily on the metaphoricity and phraseology of their metaphoric language. He found that the metaphor language performance of the L2 learners with different grade levels and writing scores was significantly different.

The individual differences mentioned above are basically related to language-related factors. Of course, some cognitive factors have also attracted the attention of researchers. For example, Johnson and Rosano (1993) explored the effects of the cognitive style and language proficiency on the comprehension of metaphor by both English native speakers and L2 learners. They found that for L2 learners, metaphor fluency was positively correlated with English communication proficiency, while field-independent cognitive style was negatively correlated with metaphor fluency and communicative proficiency. In other words, the cognitive style of field independence may inhibit the interpretation of metaphor, and L2 proficiency promotes metaphoric comprehension in L2 learners. Littlemore (2001) proposed the notion of metaphoric intelligence, emphasizing individual differences in ways of thinking in figurative language comprehension. Some people have a ‘literal’ thinking style, while others seem more capable of making metaphoric analogies.

It can be seen that the research on the influencing factors of L2 metaphor comprehension mainly focuses on the influence of language- and culture-related factors. Little attention has been paid to individual differences in cognitive ability. Therefore, the current study aimed to directly manipulate language-related executive control ability to explore its impact on L2 metaphor comprehension. We would use the Stroop task to measure language-related executive control ability. Meanwhile, the familiarity of metaphor was also a critical factor in the present study. This way, the interaction of executive control ability and metaphor familiarity could be investigated.

According to the Predication Model, we speculate that participants with high executive control ability can inhibit irrelevant connections more accurately and quickly in the CI process, so CI will need less rounds of spreading activation. Accordingly, the following hypotheses were made. First, the reaction time (RT) in metaphor comprehension is significantly different for L2 learners with different executive control ability. That is, compared with the participants with lower executive control ability, those with higher executive control ability process metaphor much faster, and with much higher accuracy. Second, the executive control ability is more influential for unfamiliar metaphor. That is, compared with the familiar metaphor, the difference in RT and accuracy for unfamiliar metaphor is much larger between L2 learners with higher and lower executive control ability.

Methods

Participants

In total, 36 college students from Beijing Normal University were recruited for the study (26 female, with an average age of 21.87 ± 1.85 years, ranging from 18 to 25 years). All participants were right-handed Chinese native speakers with normal or corrected to normal vision and normal color vision. They had learned English for more than eight years. The study was approved by the ethics committee of the School of Psychology, Beijing Normal University. All participants signed the written informed consent and were paid for their participation.

The participants were first recruited based on their CET-6 (College English Test Band 6) or TEM-4 (Test for English Majors-Band 4) scores. Then their proficiency was measured by self-assessment and language proficiency questionnaire. CET-6 is the English examination for college students of non-English majors conducted by the Ministry of Education of China. It assesses participants’ listening, reading and writing performance. The maximum score for the CET-6 is 710, and the cutoff point for success and failure is 425. The CET-6 scores for the participants not majoring in English were higher than 550, and reached an average score of 573. Besides, TEM-4 is the English examination for college students majoring in English conducted by the Ministry of Education of China. The exam assesses listening, writing, grammar, vocabulary and reading performance. The maximum score for TEM-4 is 100, and the cutoff point for success and failure is 60. The TEM-4 scores for the participants majoring in English were higher than 70, and reached an average score of 75.

In addition, participants were asked to rate their English proficiency (1 = low proficiency, 5 = quite proficient) on a 5-point scale. Furthermore, the Oxford Placement Test (OPT) was used to provide an objective assessment of participants’ English proficiency. The OPT test consists of 25 multiple choice questions and a cloze test with a total score of 50. The mean OPT score of all the participants was 41. With reference to previous studies, our participants had an intermediate level of English proficiency (Zhou et al. 2016). Table 1 shows the age of the participants, age of English acquisition, English proficiency ratings, and OPT scores.

Table 1.

Mean age (standard deviation), age of English acquisition (AoA), English proficiency ratings and Oxford Placement Test (OPT) scores.

Age	AoA	Self-rating English proficiency				OPT score
Age	AoA	Listening	Speaking	Reading	Writing	OPT score
21.79 (1.85)	8.55 (2.64)	4.15 (0.61)	4.03 (0.59)	4.13 (0.67)	3.82 (0.65)	41.22（1.21）

Metaphor sentences

Stimulus materials took the classic metaphor form in English, namely A is P, as shown in Table 2. The materials were mainly derived from previous studies (Chiappe & Chiappe, 2007; Glucksberg, McGlone, & Manfredi, 1997), in which the length of sentence was controlled, generally among four to nine words. The syntactic structure was congruous among the materials, and the properties of the source word and the target word were consistent.

Table 2.

Example of experimental materials.

Sentence type	Example
Familiar metaphor	Health is wealth
Unfamiliar metaphor	A lie is a dagger
Literal	Roses are flowers

Thirty participants who did not participate in the formal experiment and had the same background as those who did were asked to assess the familiarity of the materials (i.e. whether the metaphoric sentences are commonly known: 1 = very unfamiliar; 5 = very familiar). Finally, we selected 25 familiar metaphoric sentences, 25 unfamiliar metaphoric sentences and 25 literal sentences. The average familiarity of the literal sentences， familiar metaphoric sentences and unfamiliar metaphoric sentences was 4.67 ± 0.40, 4.49 ± 0.37, and 2.74 ± 0.37 respectively. There was significant difference in familiarity among the three types of sentences, F(2,72) = 196.25, p < 0.01. Specifically, the literal and familiar metaphoric sentences were not significantly different in familiarity, t(48) = 1.75, p > 0.05. However, the literal and unfamiliar metaphoric sentences were significantly different, t(48) = 17.81, p < 0.001. Also, the difference in familiarity between the familiar and the unfamiliar metaphoric sentences was significant t(48) = 13.92, p < 0.001.

The metaphoric words were given to the participants for preview before the experiment to exclude the impact of vocabulary on metaphor comprehension

Assessment of executive control ability

In this study, the modified Stroop task (Heidlmayr et al., 2014) was used to measure language-related execution control. There were three conditions: neutral, congruent, and incongruent stimuli. The congruent stimuli consisted of three color words (红，黄，蓝) (‘red’, ‘yellow’, ‘blue’ written in Chinese) written in the same color in which the stimulus was presented (e.g. the word 红 (‘red’) written in red ink). The incongruent stimuli consisted of the same three words with the display color not matching the word meaning (e.g. the word 蓝 (‘blue’) written in red ink). In the neutral condition, non-color meaning words (杯) (‘cup’ written in Chinese) were presented in one of the three colors. There was a total of 120 trials, including 90 (75%) consistent trials, 15 (12.5%) inconsistent trials, and 15 neutral trials (12.5%). All trials were presented pseudo-randomly using E-Prime 1.1 software. Stimuli were presented in the manner described in the study by Heidlmayr et al. (2014). First, a fixation cross was presented in the center of the screen for 500–1000 milliseconds (ms), and then the stimulus was presented for 1500 ms, and finally a blank screen was presented for 500 ms. The participants were asked to judge the color of the stimulus as quickly and accurately as possible. RT and accuracy were recorded by pressing the ‘1’ key (for ‘red’ response), the ‘2’ key (for ‘yellow’ response), or the ‘3’ key (for ‘blue’ response) with the index, middle and ring fingers of their left or right hand. The task consisted of two parts, a practice session and the formal experiment. The practice session consisted of 24 trials, including three neutral conditions, three inconsistent conditions and 18 consistent conditions. In the formal experiment, the order of stimulus presentation was controlled in a pseudo-random manner to avoid the same experimental condition appearing three times in succession and to avoid the same word or the same color appearing consecutively. Pseudo-random settings were produced with the Conan (Nowagk, 1998) program. The executive control capacity was measured by subtracting neutral condition from the inconsistent condition (i.e., the Stroop effect). The smaller the Stroop effect, the greater the executive control ability of the participant.

Procedure

The paradigm adopted in this task was a revision of the research paradigm in Benedek et al. (2014), which was programmed using E-prime 1.1 software. Instructions and practice sessions were provided to the participants before the formal experiment. All the sentences were randomly presented. First, a white fixation cross of 500 ms was presented in the center of a black screen, and then the first half of the sentence (A is) was presented for 1500 ms and finally the second half of the sentence (P) for 3000 ms. The participants were required to respond upon the second half of the sentence as quickly and accurately as possible to determine whether the sentence was metaphoric or not by pressing ‘J (for yes)’ or ‘F (for no)’. The assignment of the response keys was balanced across participants.

Results

In the metaphor comprehension task, the data of error response (8.8%) and RTs beyond ±3 standard deviation of the average mean (1.6%) were excluded. A mixed-effects logistic model of accuracy and a mixed-effects model of RT were built separately to analyze the performance of the participants (Baayen, Davidson, & Bates, 2008; Jaeger, 2008). One important reason for using the mixed effects model instead of traditional statistics is that it allows us to investigate item-related factors and continuous variables that are based on participant-related differences (e.g. executive control ability). This kind of analysis cannot be easily accomplished using traditional analysis of variance (Baayen et al., 2008). All statistical analyses were carried out using R 2.15.2 (R Core Team, 2013), implemented with package lme4 (Bates, Maechler, Bolker, & Walker, 2015), lmerTest (Kuznetsova, Brockhoff, & Christensen, 2016), and language R (Baayen, 2013). The RT and mean accuracy are presented in Table 3. The executive control ability of participants was split by the median. Half of the participants were assigned to the group with higher executive control ability, and the other half to the group with lower executive control ability.

Table 3.

Mean reaction times (RTs) (milliseconds) (standard deviations) and accuracy rates (%) of the three sentence types.

Subject	RT			Accuracy
	Familiar metaphors	Unfamiliar metaphors	Literal	Familiar metaphors	Unfamiliar metaphors	Literal
High- executive control (EC)	906 (236)	1002 (261)	876 (238)	92 (27)	87 (33)	90 (30)
Low- executive control (EC)	981 (282)	1066 (299)	915 (259)	93 (26)	90 (30)	93 (25)

For the RT data, the sentence type and the Stroop effect were fixed factors, and the participant and item were random factors. The analysis results are presented in Table 4. There was no difference in RT between familiar metaphor and literal sentences (t = 1.19, p = 0.24), but the RT for unfamiliar metaphor was significantly longer than that for the literal sentences (t = 5.17, p < 0.001). Furthermore, in order to find out the RT difference between familiar and unfamiliar metaphors sentences, we reanalyzed the data with familiar metaphor as the baseline. The result showed that the RT for unfamiliar metaphors was significantly longer than that of familiar metaphoric sentences (t = 3.90, p < 0.001). A significant interaction between the Stroop effect and the RT of familiar metaphoric sentences was found (t = 2.48, p < 0.05). A further simple effect analysis showed significant Stroop effects in familiar metaphoric sentences (familiar: t = 2.30, p < 0.05), indicating that participants with higher executive control ability comprehended familiar metaphors much faster than those with lower executive control ability. The interaction between the Stroop effect and unfamiliar metaphoric sentences was not significant (t = 0.77, p > 0.05).

Table 4.

Mixed-effects model of reaction time in the metaphor comprehension task.

Predictor	Estimate	Standard error	t value	p(\|t\|)
Fixed effects
(Intercept)	895.34	28.10	31.86	0.00
Familiar	30.86	25.85	1.19	0.24
Unfamiliar	134.38	25.99	5.17	0.00***
Stroop	0.32	0.28	1.13	0.27
Familiar: Stroop	0.35	0.14	2.48	0.01*
Unfamiliar: Stroop	0.11	0.14	0.77	0.44

p < 0.05; ***p < 0.001.

For the accuracy data, the sentence type and Stroop effect were fixed factors, and the participant and item were random factors. The analysis results are presented in Table 5. The accuracy for familiar metaphoric sentences was the same as that of literal ones (z = 0.08, p = 0.94), but the accuracy for unfamiliar metaphoric sentences was significantly lower than that of literal sentences (z = −2.00, p < 0.05). Moreover, in order to find out the difference in accuracy between familiar and unfamiliar metaphoric sentences, we reanalyzed the data with familiar metaphor as the baseline. The result showed that the accuracy for unfamiliar metaphors was significantly lower than that of familiar metaphoric sentences (z = −2.01, p < 0.05). No other significant interactions were found.

Table 5.

Mixed-effects logistic model of accuracy in the metaphor comprehension task.

Predictor	Estimate	Standard error	z value	p(\|z\|)
Fixed effects
(Intercept)	3.03	0.26	11.49	0.00
Familiar	0.02	0.31	0.08	0.94
Unfamiliar	−0.61	0.31	−2.00	0.046*
Stroop	−0.00	0.002	−0.84	0.40
Familiar: Stroop	−0.00	0.002	−1.09	0.28
Unfamiliar: Stroop	−0.00	0.002	−0.40	0.69

p < 0.05.

Discussion

The current study examined the effect of L2 learners’ executive control ability on L2 metaphor comprehension. The results showed that the executive control ability could affect L2 metaphor comprehension. Moreover, the effect of executive control ability on L2 metaphor comprehension is modulated by the familiarity of the metaphor. Specifically, for familiar metaphor sentences, the response time of participants with higher executive control ability was significantly faster than those with lower executive control ability. However, for the unfamiliar metaphor sentences and literal sentences, the effect of executive control ability was not significant.

Firstly, the current study found that L2 metaphor comprehension is related to the familiarity of metaphors. Compared with the literal sentences, the unfamiliar metaphoric sentence comprehension requires a longer RT with lower accuracy. This is consistent with previous studies on native language speakers (Glucksberg, 2001; Kintsch, 2000, 2001, etc.). According to the Predication Model, the reason may be that the number of semantic neighbors required for the comprehension of the three sentence types is not the same. Just like Kintsch (2001) says, for most sentences combining familiar terms in expected ways (i.e. literal sentences), m = 20 works well, because terms related to A will be found even among the closest neighbors of P. For metaphors where the predicate and argument can be quite distant, the crucial terms are usually not found among the top 100 neighbors, and m needs to be set larger, say 500 neighbors, or even more. Thus, the RT required is longer and the accuracy is lower for unfamiliar metaphors. However, there is no significant difference in the processing time and the accuracy rate between familiar metaphoric sentences and literal sentences, which may be due to the material setting. Take the familiar metaphor ‘Health is wealth’ as an example. With the increased frequency of use, the figurative meaning may become lexicalized over time (Bowdle & Gentner, 2005), which is similar to the literal meaning of the sentence.

Secondly, the current study further tested the effect of the executive control ability on L2 metaphor comprehension. Compared with the participants with lower executive control ability, the participants with higher executive control ability took shorter time to comprehend familiar metaphors, indicating that executive control does enhance L2 familiar metaphor comprehension. As mentioned above, the size of the semantic neighborhood of P needing to be activated in the comprehension of metaphor is larger than that of literal meaning, so a larger number of semantic neighbors which are attributes of P has to be suppressed, but not attributes of A in this process. Previous studies have shown that individuals with high WM span have more cognitive resources to devote to inhibitory tasks than individuals with low WM span. Now we believe that the factor that plays the central role is the individual difference in the executive control ability. That is to say, individuals with higher executive control ability are more accurate and faster in suppressing irrelevant information in the CI process. To be specific, those with high executive control ability are more accurate and rapid in inhibiting irrelevant links during CI, and consequently, CI will require fewer rounds of spreading activation to settle, allowing them to arrive at better interpretations faster than those with low executive control ability. For example, to comprehend the metaphor “My lawyer is a shark,” L2 learners with higher executive control ability can inhibit irrelevant properties of sharks more quickly and accurately during CI. Therefore, CI only needs less rounds of spreading activation to complete metaphor comprehension. As a result, they are more efficient and accurate in L2 metaphor comprehension. To conclude, the results of this study have provided evidence for the explanatory power of the Predication Model on L2 learners about the effect of the executive control ability on L2 metaphor comprehension.

Thirdly, this study also found that L2 unfamiliar metaphor comprehension in L2 learners was not affected by the executive control ability. The results of unfamiliar metaphors are not consistent with our hypothesis. We expected that participants with higher executive control ability would process L2 unfamiliar metaphors more quickly and more accurately. We also expected that the differences in RT and accuracy for the unfamiliar metaphoric sentences between participants with high and low executive control ability would be bigger than that for familiar metaphoric sentences. However, we found that the interaction between the unfamiliar metaphoric sentence and the executive control was not significant. Does this mean that L2 learners’ executive control ability has no effect on unfamiliar metaphoric sentence comprehension? Maybe this is not the case. The possible reasons are as follows: first, the difficulty of the sentences may cause the emergence of the floor effect. That is to say, the difficulty of the unfamiliar metaphor sentence is so high that the participants either with high or low executive control ability find it too difficult to process unfamiliar metaphors. Chiappe and Chiappe (2007) found that participants with higher WM and better inhibition control in the Stroop task were faster in constructing metaphor interpretations and had higher quality interpretations, which has nothing to do with the difficulty of metaphor. However, we find that their study divides the difficulty of metaphor through topic restriction and vehicle ambiguity. Metaphors with low ambiguity (e.g. some jobs are prisons) vehicles and those with high-constraint (e.g. Jealousy is an infection) topics were regarded as easier to comprehend, and those with low-constraint (e.g. some people are puzzles) topics and those with high-ambiguity (e.g. some colleges are gardens) vehicles were regarded as more difficult. The whole distance of metaphoric difficulty may be too small, and the metaphoric sentences may not be too difficult overall, so there would be no difference in their results. Moreover, even for the same metaphor, the difficulty might increase for L2 learners (Kecskes, 2006; Littlemore et al., 2011). Therefore, the role of familiarity for L2 metaphor comprehension may not be the same as the L1 metaphor. Second, it may be the effect of the similarity of concept knowledge between L1 and L2 metaphor. In our material, the concept of knowledge for the familiar metaphor materials selected are similar between the native language and the L2 (e.g. Zeal is fire), but this is not the case for some of the unfamiliar metaphors selected (e.g. my sister is a dragon). Türker (2016) found that L2 learners processed metaphors better when L1 and L2 have similarities. Therefore, greater difficulty would be met by the participants when processing L2 unfamiliar metaphor, further blocking the effect of executive control. Such a result can still be explained by the Predication Model. According to the Predication Model, the size of semantic neighborhood of P that needs to be activated in the comprehension of familiar metaphors is more than that for literal sentences. It may be even bigger for unfamiliar metaphors, and in the process of CI, more rounds of diffusion activation will be needed to promote unfamiliar comprehension, so the individual’s executive control ability required will be higher. In addition, unfamiliar metaphoric sentences will be more difficult to comprehend in the L2, making it far beyond the ability of the participants to inhibit irrelevant information, regardless of their executive control ability.

To sum up, there are both similarities and differences for the role of executive control ability on L2 and L1 metaphor comprehension. We think there may be several reasons: firstly, the basic mechanisms of L1 and L2 metaphor comprehension are the same, such as the need to enhance the activation of related attributes, and inhibit the activation of irrelevant attributes. Secondly, the mechanisms of L2 metaphor comprehension may be more complex. As mentioned earlier, there are some differences in metaphor comprehension between native speakers and L2 learners, such as cultural background, sociolinguistic experience, and dominant hemisphere of processing. These factors, as an intervening variable in the process of comprehension, will influence L2 metaphor comprehension. Finally, the inhibition mechanism of L2 metaphor comprehension may be different. L2 learners’ not only need to suppress the same irrelevant information as native speakers, but also need to suppress the influence of factors such as different language experiences, and cultural background, in the process of metaphor comprehension, which lead to the increased demands on executive control. We believe that an important consequence of all these differences is that they lead to an increase in the recruitment of cognitive resource and executive control for L2 learners when comprehending L2 metaphors, reflecting differences in comprehension difficulty. This can also be explained by the Predication Model. That is, it is possible that the semantic neighbors of P in the L2 metaphor (A is P) are much further and less in the semantic network than in the native language for the same A, so its processing requires the participation of more cognitive resources.

We did not balance the difficulty of the materials, like some previous studies (Faust & Mashal, 2007; Subramaniam, Beeman, Faust, & Mashal, 2013; Subramaniam, Faust, Beeman, & Mashal, 2012). Because there is some connection between the difficulty and familiarity of metaphor, for example, some studies directly define familiar metaphors as familiar and easy to interpret, and unfamiliar metaphors as unfamiliar and difficult to interpret (Lai & Curran, 2013; Lai, Curran, & Menn, 2009).

In conclusion, this study explored the impact of the language-related executive control ability of L2 learners on metaphor comprehension by directly manipulating executive control ability. The results suggested that executive control ability of L2 learners has a significant effect on L2 metaphor comprehension, especially for the familiar L2 metaphor. It indicates that L2 learners’ executive control ability enhances their comprehension of L2 metaphors. The results have specified the effect of individual differences in executive control ability on metaphor comprehension in L2 learners, supporting the Predication Model and extending this view to L2 learners.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author biographies

Junmei Lü is a PhD student from the School of Psychology at Beijing Normal University. Her research interests focus on the processing of metaphor.

Lijuan Liang works as a Lecturer at the School of English and Education at Guangdong University of Foreign Studies. Her research focuses on phonological-orthographic processing during second language acquisition.

Baoguo Chen is a Professor from the School of Psychology at Beijing Normal University. He mainly conducts research on L2 vocabulary learning, L2 semantic and syntactic processing.

References

Baayen

R. H.

(2013). LanguageR: Data sets and functions with “Analyzing Linguistic Data: A practical introduction to statistics.” R package version 1.4. Retrieved from http://CRAN.R-project.org/package=languageR

Baayen

R. H.

Davidson

D. J.

Bates

D. M.

(2008). Mixed-effects modeling with crossed random effects for subjects and items. Journal of Memory and Language, 59(4), 390–412.

Bates

Maechler

Bolker

Walker

(2015). lme4: Linear mixed-effects models using ‘Eigen’ and S4. R package version 1.1–13. Retrieved from http://CRAN.R-project.org/package=lme4

Benedek

Beaty

Jauk

Koschutnig

Fink

Silvia

P. J.

. . .Neubauer

A. C.

(2014). Creating metaphors: The neural basis of figurative language production. NeuroImage, 90(100), 99–106.

Blanco-Elorrieta

Pylkkänen

(2016). Bilingual language control in perception versus action: MEG reveals comprehension control mechanisms in anterior cingulate cortex and domain-general control of production in dorsolateral prefrontal cortex. The Journal of Neuroscience, 36(2), 290–301.

Blasko

D. G.

Connine

C. M.

(1993). Effects of familiarity and aptness on metaphor processing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 19(2), 295–308.

Bowdle

B. F.

Gentner

(2005). The career of metaphor. Psychological Review, 112(1), 193–216.

Branzi

F. M.

Calabria

Boscarino

M. L.

Costa

(2016). On the overlap between bilingual language control and domain-general executive control. Acta Psychologica, 166, 21–30.

Carriedo

Corral

Montoro

P. R.

Herrero

Ballestrino

Sebastián

(2016). The development of metaphor comprehension and its relationship with relational verbal reasoning and executive function. PLoS One, 11(3), e0150289.

10.

Charteris-Black

(2002). Second language figurative proficiency: A comparative study of Malay and English. Applied Linguistics, 23(1), 104–133.

11.

Cheung

W. S.

(2013). Cantonese–English bilinguals’ L2 metaphor comprehension. Master of Arts in Linguistics, The Chinese University of Hong Kong.

12.

Chiappe

D. L.

Chiappe

(2007). The role of working memory in metaphor production and comprehension. Journal of Memory and Language, 56(2), 172–188.

13.

Columbus

Sheikh

N. A.

Côté-Lecaldare

Häuser

Baum

S. R.

Titone

(2015). Individual differences in executive control relate to metaphor processing: An eye movement study of sentence reading. Frontiers in Human Neuroscience, 8, 1057.

14.

Faust

Mashal

(2007). The role of the right cerebral hemisphere in processing novel metaphoric expressions taken from poetry: A divided visual field study. Neuropsychologia, 45(4), 860–870.

15.

Ferreira

L. C.

(2008). A psycholinguistic study on metaphor comprehension in a foreign language. ReVEL, 6(11), 1–23.

16.

Glucksberg

(2001). Understanding figurative language: From metaphor to idioms. Oxford, UK: Oxford University Press.

17.

Glucksberg

McGlone

M. S.

Manfredi

(1997). Property attribution in metaphor comprehension. Journal of Memory and Language, 36(1), 50–67.

18.

Goldstein

Arzouan

Faust

(2012). Killing a novel metaphor and reviving a dead one: ERP correlates of metaphor conventionalization. Brain and Language, 123(2), 137–142.

19.

Heidlmayr

Moutier

Hemforth

Courtin

Tanzmeister

Isel

(2014). Successive bilingualism and executive functions: The effect of second language use on inhibitory control in a behavioural Stroop colour word task. Bilingualism: Language and Cognition, 17(3), 630–645.

20.

Hoàng

Đ. H. T

. (2015). Metaphorical language in second language learners’ essays: Products and processes. Doctor of Philosophy in Applied Linguistics, Victoria University of Wellington.

21.

Jaeger

T. F.

(2008). Categorical data analysis: Away from ANOVAs (transformation or not) and towards logit mixed models. Journal of Memory and Language, 59(4), 434–446.

22.

Jin

(2011). Second language acquisition of spatial metaphors in English and Chinese writings: Insights from native and learner language corpora. Doctor of philosophy, The University of Arizona.

23.

Johnson

Rosano

(1993). Relation of cognitive style to metaphor interpretation and second language proficiency. Applied Psycholinguistics, 14(2), 159–175.

24.

Kazmerski

V. A.

Blasko

D. G.

Dessalegn

B. G.

(2003). ERP and behavioral evidence of individual differences in metaphor comprehension. Memory and Cognition, 31(5), 673–689.

25.

Kecskes

(2006). On my mind: Thoughts about salience, context and figurative language from a second language perspective. Second Language Research, 22(2), 219–237.

26.

Kintsch

(2001). Predication. Cognitive Science, 25(2), 173–202.

27.

Kintsch

(2000). Metaphor comprehension: A computational theory. Psychonomic Bulletin and Review, 7(2), 257–266.

28.

Kuznetsova

Brockhoff

P. B.

Christensen

R. H. B.

(2016). lmerTest: Tests in Linear Mixed Effects Models. R package version 2.0–33. Retrieved from http://CRAN.R-project.org/package=lmerTest

29.

Lai

V. T.

Curran

(2013). ERP evidence for conceptual mappings and comparison processes during the comprehension of conventional and novel metaphors. Brain and Language, 127(3), 484–496.

30.

Lai

V. T.

Curran

Menn

(2009). Comprehending conventional and novel metaphors: An ERP study. Brain Research, 1284, 145–155.

31.

Littlemore

(2001). Metaphoric intelligence and foreign language learning. Humanising Language Teaching Magazine, 3(2), Retrieved from http://www.hltmag.co.uk/mar01/mart1.htm

32.

Littlemore

Chen

P. T.

Koester

Barnden

(2011). Difficulties in metaphor comprehension faced by international students whose first language is not English. Applied Linguistics, 32(4), 408–429.

33.

Magezi

D. A.

Khateb

Mouthon

Spierer

Annoni

J. M.

(2012). Cognitive control of language production in bilinguals involves a partly independent process with-in the domain-general cognitive control network: Evidence from task-switching and electrical brain activity. Brain and Language, 122(1), 55–63.

34.

Mashal

(2013). The role of working memory in the comprehension of unfamiliar and familiar metaphors. Language and Cognition, 5(4), 409–436.

35.

Mashal

Borodkin

Maliniak

Faust

(2015). Hemispheric involvement in native and non-native comprehension of conventional metaphors. Journal of Neurolinguistics, 35, 96–108.

36.

Mashal

Faust

(2009). Conventionalization of novel metaphors: A shift in hemispheric asymmetry. Laterality, 14(6), 573–589.

37.

Mashal

Faust

Hendler

Jung-Beeman

(2007). An fMRI investigation of the neural correlates underlying the processing of novel metaphoric expressions. Brain and Language, 100(2), 115–126.

38.

Mashal

Vishne

Laor

Titone

(2013). Enhanced left frontal involvement during novel metaphor comprehension in schizophrenia: Evidence from functional neuroimaging. Brain and Language, 124(1), 66–74.

39.

Mercier

Pivneva

Titone

(2014). Individual differences in inhibitory control relate to bilingual spoken word processing. Bilingualism: Language and Cognition, 17(1), 89–117.

40.

Nowagk

(1998). Conan: A Barbarian tool for constrained randomization (Version 1.9). Leipzig: Max-Planck-Institute of Cognitive Neuroscience Leipzig.

41.

Olkoniemi

Ranta

Kaakinen

J. K.

(2016). Individual differences in the processing of written sarcasm and metaphor: Evidence from eye movements. Journal of Experimental Psychology: Learning, Memory, and Cognition, 42(3), 433–450.

42.

Pierce

R. S.

MacLaren

Chiappe

D. L.

(2010). The role of working memory in the metaphor interference effect. Psychonomic Bulletin and Review, 17(3), 400–404.

43.

Safarnejad

Ho-Abdullah

Awal

N. M.

(2014). A cognitive study of happiness metaphors in Persian and English. Procedia-Social and Behavioral Sciences, 118, 110–117.

44.

Sandgren

J. M. S.

(2014). Problems L2 readers encounter with metaphor in discipline specific texts. Doctor of philosophy, University of Wellington.

45.

St Clair-Thompson

H. L.

Gathercole

S. E

. (2006). Executive functions and achievements in school: Shifting, updating, inhibition, and working memory. The Quarterly Journal of Experimental Psychology, 59(4), 745–759.

46.

Subramaniam

Beeman

Faust

Mashal

(2013). Positively valenced stimuli facilitate creative novel metaphoric processes by enhancing medial prefrontal cortical activation. Frontiers in Psychology, 4, 211.

47.

Subramaniam

Faust

Beeman

Mashal

(2012). The repetition paradigm: Enhancement of novel metaphors and suppression of conventional metaphors in the left inferior parietal lobe. Neuropsychologia, 50(12), 2705–2719.

48.

R Core Team. (2013). R: A language and environment for statistical computing. Vienna and Austria: R Foundation for Statistical Computing. Retrieved from http://www.R-project.org/

49.

Türker

(2016). The role of L1 conceptual and linguistic knowledge and frequency in the acquisition of L2 metaphorical expressions. Second Language Research, 32(1), 25–48.

50.

Verwoerd

Wessel

de Jong

P. J.

(2009). Individual differences in experiencing intrusive memories: The role of the ability to resist proactive interference. Journal of Behavior Therapy and Experimental Psychiatry, 40(2), 189–201.

51.

Zhou

Rossi

Liu

Chen

(2016). Effects of working memory capacity in processing wh-extractions: Eye-movement evidence from Chinese–English bilinguals. Journal of Research in Reading. doi:10.1111/1467-9817.12079.