Content uploaded by Bálint Forgács
Author content
All content in this area was uploaded by Bálint Forgács on Sep 20, 2020
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
Integrative Systems
An Electrophysiological Abstractness Effect for
Metaphorical Meaning Making
Bálint Forgács
1,2
https://doi.org/10.1523/ENEURO.0052-20.2020
1
Laboratoire Psychologie de la Perception (LPP), Université de Paris, Paris 75006, France and
2
MTA-ELTE Social
Minds Research Group, Eötvös Loránd University (ELTE), Budapest 1064, Hungary
Abstract
Neuroimaging studies show that metaphors activate sensorimotor areas. These findings were interpreted as
metaphors contributing to conceptual thought by mapping concrete, somatosensory information onto abstract
ideas. But is sensorimotor information a necessary constituent of figurative meaning? The present study em-
ployed event-related potentials (ERPs) in a divided visual field paradigm with healthy adults to explore the role
of sensorimotor feature processing in the comprehension of novel metaphors via the electrophysiological con-
creteness effect. Participants read French, novel adjective-noun expressions that were either metaphorical
(“fat sentence”) or literal (“fat hip”). While literal expressions evoked a typical concreteness effect, an en-
hanced frontal negativity during right hemisphere (RH) as opposed to left hemisphere (LH) presentation, meta-
phors showed no such sign of sensorimotor feature processing. Relative to literals, they evoked a sustained
frontal negativity during LH presentation and similar amplitudes during RH presentation, but both of these ef-
fects were the greater the more abstract the metaphors were. It is the first time such an electrophysiological
abstractness effect is reported, just the opposite of a concreteness effect. It is particularly noteworthy that
ERPs evoked by metaphors were not contingent on figurativeness, novelty, meaningfulness, imageability,
emotional valence, or arousal, only on abstractness. When compared with similarly novel literal expressions,
metaphors did not evoke a typical N400 and did not activate the RH either. The findings shed new light on the
neurocognitive machinery of figurative meaning construction, pervasive in everyday communication. Contrary
to embodied cognition, the conceptual system might be organized around abstract representations and not
sensorimotor information, even for lush, metaphorical language.
Key words: concreteness effect; EEG; embodiment; figurative language; metaphor; novel language
Significance Statement
In the past decades, several popular theories have been promoting the idea that the format of semantic rep-
resentation is sensomotoric and/or based on experience. The abstractness effect reported here challenges
such Empiricist accounts of the conceptual system, like embodiment or connectionism. It argues against
metaphors being the vehicles of transmitting sensorimotor information toward higher domains of cognition.
Instead, even perceptual metaphors appear to evoke neural responses contingent on their abstractness.
The current findings also challenge long held notions about the central role of the right cerebral hemisphere
or literal meaning in comprehending figures of speech and suggest instead that the brain is not sensitive to
figurativeness per se but to the abstractness it brings along.
Received February 9, 2020; accepted July 13, 2020; First published August 17,
2020.
The author declares no competing financial interests.
Author contributions: B.F. designed research; B.F. performed research; B.F.
analyzed data; B.F. wrote the paper.
September/October 2020, 7(5) ENEURO.0052-20.2020 1–12
Research Article: New Research
Introduction
Metaphor has been a small but crucial area of study in
the cognitive neuroscience of language, and some
scholars consider it to be an essential feat of human
cognition (Lakoff and Johnson, 1980;Mithen, 1996;
Pinker, 2010). The aim of the current study was to test
the role of literal meaning, or sensorimotor processes,
as embodiment frames it (Lakoff and Johnson, 1999), in
their comprehension.
Metaphors are often described as vivid and imagistic
conceptual tools and have been proposed to rely on
sensorimotor computations (Gallese and Lakoff, 2005;
Lakoff, 2014). Several functional magnetic resonance
imaging (fMRI) studies found that metaphors activate
sensorimotor brain regions for motion (Boulenger et al.,
2009,2012), texture (Laceyetal.,2012), taste (Citron
and Goldberg, 2014), smell (Pomp et al., 2018), and mo-
tion sensitive visual areas or their vicinity for action
(Chen et al., 2008;Saygin et al., 2010;Desai et al.,
2011;Cardillo et al., 2012;Lacey et al., 2017). However,
as Mahon and Caramazza (2008) pointed out, the acti-
vation of sensorimotor areas could be because of auto-
matic spreading activation, which could have driven
fMRI results without any causal role.
It has been argued that sensorimotor processes play a
critical role in word comprehension because of their early in-
volvement (;200 ms; Pulvermüller, 2005;Pulvermüller et al.,
2005). Early automatic activations of alternative meanings of
ambiguous words, however, are suppressed by 200–
300 ms (Swinney, 1979;Seidenberg et al., 1982;Gergely
and Pléh, 1994), just as motor activations for action words
(Shtyrov et al., 2014). Therefore, literal senses of metaphors
and corresponding sensorimotor activations could be sup-
pressed during lexical access. Schneider et al. (2014)’s
metaphor study found a single connection between event-
related potentials (ERPs) and BOLD signals in an early, 200-
to 270-ms time window, which raise the possibility that fMRI
studies may not have reported processes beyond early (au-
tomatic, spreading) sensorimotor activations for figurative
meaning.
Another indicator of sensorimotor processes is the electro-
physiological concreteness effect. It can be evoked by con-
crete (e.g., “chair”) as opposed to abstract words (e.g.,
“truth”) and has two negative frontal components: the con-
creteness N400, reflecting the activation of a larger set of per-
ceptual semantic features, and the N700, thought to indicate
mental imagery (Kounios and Holcomb, 1994;Holcomb et
al., 1999;West and Holcomb, 2000), specific to the right
hemisphere (RH) (Huang et al., 2010). Barber et al. (2013)
showed that concreteness effects can be evoked even when
concrete and abstract words are matched in terms of image-
ability and semantic richness: they propose that the frontal
N400 reflects the activation of multimodal features, while the
N700 is related to the integration of such sensorimotor fea-
tures into a representation.
Only a couple of studies looked into the sensorimotor
processing of metaphors. Two of them reported an en-
hanced N400 response, an indicator of demanding se-
mantic processing (Kutas and Federmeier, 2011;Brouwer
et al., 2012). However, one did not balance stimuli for fa-
miliarity and imageability (Schmidt-Snoek et al., 2015),
while the other did not match metaphorical and literal
conditions in terms of novelty (Shen et al., 2015), thus the
reported effects might be driven by novelty instead of fig-
urativeness. One study attempted to control for novelty
employing target nouns that combined into unfamiliar
metaphorical, and concrete and abstract literal expres-
sions (Forgács et al., 2015). The authors found that only
less concrete (i.e. abstract) metaphors evoked a greater
negativity in the N400 time window than abstract expres-
sions, more concrete metaphors, paradoxically, did not.
Several groups reported typical centro-parietal N400
effects for novel metaphors, and interpreted them as re-
flecting mappings (Coulson and Van Petten, 2002;Rutter
et al., 2012;Lai and Curran, 2013;Schneider et al., 2014;
Rataj et al., 2018) or as the activation of literal senses
(Pynte et al., 1996;Tartter et al., 2002;Weiland et al.,
2014). In lack of matched novel literal conditions, how-
ever, these studies could have reported mere novelty
N400 effects (cf. Davenport and Coulson, 2013).
Thegoalofthepresentexperimentistoinvestigatethe
role of sensorimotor feature processing in the comprehen-
sion of novel perceptual metaphors via the electrophysiolog-
ical concreteness effect. The following predictions were
made.Accordingtostrongembodiment,metaphorsshould
not differ from literal expressions in terms of the frontal
N400 and N700 concreteness effects during RH presenta-
tion. If, however, metaphors are comprehended without
sensorimotor computations (besides early automatic activa-
tions), they should evoke a reduced N700 in the RH. Once
controlled for novelty, metaphors should neither evoke a
typical N400 effect relative to literal expressions nor activate
theRHrelativetotheLH.
Materials and Methods
Participants
Thirty-six healthy adults (24 female) between ages
18–35 (M = 25, SD = 4) took part in the experiment.
Participants were recruited through the RISC website
(http://www.risc.cnrs.fr)andreceivede15 compensa-
tion for their work. They were all native speakers of
French, right handed as measured by the Edinburgh
Handedness Inventory (Oldfield, 1971) scoring above
75 points (M = 92.1, SD = 9.21), had normal or corrected
to normal vision, and reported no neurologic or
This work was supported by the postdoctoral fellowships of the Fyssen
Foundation and the Hungarian Academy of Sciences (462003) and a Young
Researcher Grant (125417) of the National Research, Development and
Innovation Office of Hungary (to B.F.).
Acknowledgements: I thank the invaluable help and support of Judit Gervain
in realizing this work; Lornah Mooken, Merve Bulut, and Vincent Forma for
their ingenious help in stimulus creation; and Nausicaa Pouscoulous and
Csaba Pléh for their suggestions on the first version of this manuscript.
Correspondence should be addressed to Bálint Forgács at forgacs.
balint@ppk.elte.hu.
https://doi.org/10.1523/ENEURO.0052-20.2020
Copyright © 2020 Forgács
This is an open-access article distributed under the terms of the Creative
Commons Attribution 4.0 International license, which permits unrestricted use,
distribution and reproduction in any medium provided that the original work is
properly attributed.
Research Article: New Research 2 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
psychiatric problems. Sample size was determined
based on prior research (Lee and Federmeier, 2008;
Huang et al., 2010;Forgács et al., 2015). An additional
eight participants were excluded from statistical analy-
ses because of excessive blink, eye-movement, and
other ERP artifacts.
Stimuli
The study was conducted in French, where, as in Latin
languages, adjectives typically follow nouns, which allows
for measuring neural responses right at the adjective: the
concrete, physical vehicle of the metaphor. An initial set
of 320 word triplets were generated where a physical ad-
jective, which refers to a perceptual experience (e.g.,
“tasty”) modified two nouns to form a metaphorical (e.g.,
“tasty dependence”) and a literal expression (“tasty
plum”). Care was taken to make sure that none of the
metaphors had a possible literal interpretation. The bi-
gram frequency of metaphorical and literal word pairs
was kept below 20 hits in a Google search of the French
web to ensure that the expressions were novel, as novel
metaphors are more likely to evoke sensorimotor (Binder
and Desai, 2011) and RH processes (Bohrn et al., 2012).
Adjectives were maximum 11 characters long (M = 7.42,
SD = 1.67) to aid readability during lateralized presenta-
tion. The word pairs were rated in a norming study by 103
volunteers recruited via the RISC website, who received a
e10/h compensation, and did not take part in the EEG ex-
periment. Participants rated the expressions on seven-
point Likert scales along three randomly assigned tasks
for meaningfulness (“How meaningful is it?”), concrete-
ness (“How easy is it to experience with the senses?”),
and to avoid the definition or explanation of the notion of
metaphorical, literalness (“How literal is it?”). The best 200
word triplets (forming two expressions) were selected
with the highest meaningfulness values (at least 2.5), with
the concreteness of the metaphor being lower than that of
its literal counterpart (41 items of the original 320 showed
an inverse pattern and were excluded), and the literalness
of the metaphor being lower than that of its literal counter-
part (an additional 62 items showed the inverse and were
excluded). The final 200 items were rated in a second
norming study by 63 volunteers (with the same conditions
as above) on a seven-point Likert scale according to
imageability (“How imaginable is it?”), valence (“What is
its emotional valence?”), arousal (“How arousing is it?”),
and the concreteness of the noun (“How easy is it to ex-
perience with the senses?”). Pairwise comparisons using
paired ttests, and an independent sample ttest for noun
concreteness, revealed significant differences for all vari-
ables but bigram frequency and arousal. Metaphors were
slightly less meaningful (still around the median of the
scale), less literal, concrete, imageable, and slightly more
emotionally negative than literal counterparts. Nouns in
literal expressions were more concrete than in metaphori-
cal expressions. Norming results are presented in Table 1.
The full stimulus list can be viewed in Extended Data
Table 1-1.
Experimental procedures
Huang et al. (2010)’s word pair paradigm has been
adopted, which measured ERPs using the divided visual
field technique, to compare identical target words both in
metaphorical and literal constructs, and to avoid confound-
ing sentence processing effects. The experiment was ap-
proved by the Ethics Committee of Université de Paris, and
participants gave their written informed consent on arrival to
the lab. They were seated in a dimly lit room 60 cm from a
screen with their head placed on a chinrest. Their task was
to read word pairs (“scarified history”) followed by a probe
word (“amnesia”) and to decide, using a button box,
whether the probe was semantically related to the combined
meaning of the preceding two-word expression or not. Such
a procedure was adapted to make sure participants read for
comprehension. They read the instructions and completed
16 practice trials before the experiment started. Stimuli was
presented on a black background in 28-point Arial capital
letters. Each trial started with four plus signs “1111”
(1000 ms), after which a blank screen appeared with a jitter
(800–1200ms). Next, the prime word (noun) appeared cen-
trally for 200 ms, and after a jittered blank screen (300–
400 ms), the target word (adjective) was presented for
200 ms either to the left or right visual field. The inner most
edge of target words were 1.5° visual angle away from the
center of the screen. After a blank screen (1300ms), the
probe word appeared (200 ms), which was followed by an-
other blank screen (800ms) and a question mark “?”,which
remained on the screen until participants responded.
Table 1: Psycholinguistic properties and differences of the stimuli (N= 200)
Metaphors mean (SD) Literals mean (SD) Levene’sFvalue tvalue/ Svalue 95% confidence interval
Bigram frequency 3.81 (4.36) 4.06 (4.29) 0.293 0.652 [–0.496, 0.986]
Meaningfulness 4.04 (0.92) 5.39 (0.95) 0.009 15.8
ppp
[1.18, 1.51]
Literalness 3.37 (0.75) 5.11 (0.86) 3.92
p
200
ppp
[1.40, 1.85]
Concreteness 2.80 (0.68) 4.49 (0.92) 21.7
ppp
200
ppp
[1.44, 1.81]
Imageability 4.06 (0.66) 5.28 (0.67) 0.018 19.5
ppp
[1.09, 1.34]
Valence (63) –0.84 (1.15) –0.66 (1.27) 3.63 2.53
p
[0.038, 0.309]
Arousal 3.81 (0.83) 3.71 (0.88) 0.550 –1.87 [–0.202, 0.005]
Noun concreteness 2.49 (0.72) 4.29 (0.73) 3.89
p
191
ppp
[1.57, 1.92]
When the variance of a measure was not equal between metaphors and literals, as reported by a median centered Levene’s test (Fvalue), a sign test (Svalue)
was used, otherwise a ttest (tvalue). Metaphorical and literal expressions differed on all properties except for their bigram frequency and arousal value. The full
stimulus list is shown in Extended Data Table 1-1.
pp,0.05.
ppp p,0.001.
Research Article: New Research 3 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
Except during the question mark a small red dot was pre-
sented centrally and slightly below the words; participants
were requested to fixate it during lateralized presentation
and to try not to blink when it was visible. The experiment
took ;35–40 min with short breaks between the five blocks
of stimuli, each consisting of 40 trials. Participants were al-
lowed to take short blink breaks if necessary also when the
question mark was present. Each individual was assigned
an individual pseudo-randomized stimulus file, with no more
than three consecutive trials in either visual fields, and no
more than two consecutive word pairs from the same condi-
tion. Stimulus presentation and EEG triggering was con-
trolled via E-prime 2.08 software (Psychology Software
Tools). All raw EEG data, stimulus code, analysis scripts,
and full statistical results are available at http://osf.io/j45gt/.
EEG recording and analysis
The EEG signal was recorded continuously with EGI’s
128-channel HydroCel Geodesic Sensor Net at 500-Hz
sampling rate. Electrode impedance was kept below 50
kVand readjusted during breaks when necessary. The
EEG recording was analyzed with Net Station 4.5.6
(Electrical Geodesics Inc.). Raw EEG data were filtered
using a 0.3-Hz high-pass and a 30-Hz low-pass filter and
segmented into epochs 200 ms before and 1200 ms after
the onset of target words. Automatic artifact detection al-
gorithms for blinks, eye-movements, and bad channels
were used to reject bad segments, which was confirmed
via visual inspection. Bad channels were replaced by
spherical spline interpolation, and the data was baseline-
corrected to the 200-ms preceding word onset and re-ref-
erenced to the average reference. Participants needed to
produce at least 30 clean trials per condition per visual
field to be included in the final statistical tests. On aver-
age, participants contributed 43 trials to each of the four
conditions (86%). For the typical N400 effect, mean am-
plitudes were extracted between 300–500 ms (Kutas and
Federmeier, 2011) and averaged over a parietal region of
interest (ROI) that included the following electrodes: 31,
37, 41, 42, 46, 47, 51, 52, 53, 54, 55, 58, 59, 60, 61, 62,
65, 66, 67, 70, 71, 72, 75, 76, 77, 78, 79, 80, 83, 84, 85,
86, 87, 90, 91, 92, 93, 96, 97, 98, 102, 103, VREF (Fig. 1).
Concreteness effect ERP responses were analyzed over a
frontal ROI (that included the following electrodes: 2, 3, 4,
5, 6, 7, 9, 10, 11, 12, 13, 15, 16, 18, 19, 20, 22, 23, 24, 26,
27, 28, 29, 30, 34, 35, 36, 40, 104, 105, 106, 109, 110,
111, 112, 116, 117, 118, 123, 124) in the 300- to 500-ms
and 700- to 1000-ms time windows, following Lucas et al.
(2017), who used novel conceptual combinations to study
concreteness.
Statistical analyses
All statistical tests reported here were conducted over
single trials using linear mixed-effects modeling (LMEM:
Baayen, 2008;Baayen et al., 2008), with the statistical
language R (Core Team R, 2017) and the lme4 package
(Bates et al., 2015). Data points .2.5 SD away from each
individual’s mean ERPs were removed. The following
steps were taken during model building. First, the order of
trials was introduced as a fixed effect against a model of
random effects only, to check whether responses were
modulated by fatigue; it was included only if it significantly
improved the model. Presentation side (RH and LH) and
word pair category (metaphor and literal) were entered in
the models as fixed effects in interaction; participants and
items were entered as random effects, with random
slopes and intercepts for side, category, and their interac-
tion, to keep it maximal (Barr et al., 2013). If the random
effect structure was too complex for the model to con-
verge, it was simplified stepwise. To control for possible
confounds and specify the role of psycholinguistic varia-
bles, they were included in the statistical models as cova-
riates. They were entered separately to avoid collinearity,
as some were strongly correlated. The effect of emotional
factors was checked by adding side valence and side
arousal interaction terms only, since they were not ex-
pected to affect the two conditions differentially. Next, the
logarithm of bigram frequency and semantic covariates
(meaningfulness, concreteness, imageability and literal-
ness) were introduced one-by-one to the model extending
the side category interaction into a three-way interac-
tion. Covariates were included in the final model only if
they improved it significantly (and otherwise are not re-
ported). Models were compared using likelihood ratio
tests, and pvalues of the final models were calculated
based on the Kenward–Roger approximation with the
mixed() function (Singmann and Kellen, 2017). Likelihood
tests for model building are not reported in the main text
but are fully available at http://osf.io/j45gt/. Model resid-
ual plots did not exhibit visible deviations from normality
and homoscedasticity.
Results
Behavioral results
Response accuracy was highly variable across individu-
als but similar for related (M = 68%, SD = 47%) and unre-
lated probe words (M = 65%, SD = 48%), which suggests
that participants payed attention and made an effort at in-
terpreting the two-word expressions by linking their com-
bined meaning to the probe words.
Confirmatory ERP analyses
Frontal and parietal electrophysiological responses are
shown over two exemplar electrodes together with mean am-
plitudes in selected ROIs and time windows in Figure 1
(grand average ERPs at each electrode site can be viewed in
Extended Data Figs. 1-1,1-2 for LH and RH presentation, re-
spectively). Statistical analyses of two indicators of lateralized
processing of visual stimuli, the N1 (100–200 ms) and the se-
lection negativity (300–1000 ms), confirmed that lateralized
presentation was successful in this study (Extended Data Fig.
1-3 and in full detail at http://osf.io/j45gt/).
Typical N400
Mean amplitudes were calculated in the 300- to 500-ms
time window over electrodes in the parietal ROI. The
final model [n400.post.lmem = mixed(n400 ;trial 1cate-
gorypside 1(1 1category|participant) 1(1 1side|item_nr),
n400.post.data)] revealed no significant difference between
Research Article: New Research 4 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
the two categories (
b
=–0.07, SE = 0.04, F
(1,34.4)
= 3.37, p=
0.075) only between the two presentation sides (
b
=0.12,
SE = 0.04, F
(1,197)
=8.44, p= 0.004), with no interaction (
b
=
0.01, SE = 0.04, F
(1,5892)
=0.11, p= 0.74). Metaphors did not
evoke a typical N400 relative to literals, and both conditions
evoked a greater negativity during LH presentation than RH
presentation (Figs. 1,2). One participant only had 29 trials in
one of the bins for the typical N400 analysis, but the exclu-
sion of this individual did not change the pattern of results:
only side had a significant main effect (p= 0.002), category
did not (p= 0.11), and there was no interaction (p=0.84;
Bonferroni corrected
a
= 0.025).
Frontal N400
For the frontal N400 mean amplitudes were calculated
in the 300- to 500-ms time window over the frontal ROI
(Figs. 1,2). The final model [n400.ant.lmem = mixed(n400
;trial 1categorypside 1(1|participant) 1(1|item_nr),
n400.ant.data)] revealed a category main effect (
b
=
–0.15, SE = 0.04, F
(1,5986)
= 15.4, p,0.001) but also an in-
teraction between side and category (
b
= 0.08, SE = 0.04,
F
(1,6003)
= 4.91, p= 0.027). When the interaction was bro-
ken down by presentation side, metaphors evoked a
greater negativity than literals during LH presentation (
b
=
–0.23, SE = 0.05, F
(1,2917)
= 19.4, p,0.001) but not during
RH presentation (
b
=–0.06, SE = 0.05, F
(1,2958)
= 1.43,
p= 0.23).
Frontal N700
Over the frontal ROI mean amplitudes were calculated
between 700 and 1000 ms for the N700 (Figs. 1,2). Trial
and emotional valence significantly improved the final
Figure 1. ERP responses on a frontal (Fz) and a parietal (Pz) exemplar electrode (upper left and right panels, respectively). Blue shades in-
dicate the N400 (300–500 ms) and the N700 (700–1000ms), the green shade the N400 time window. Negative is plotted upwards. Lower
panels show bar charts of amplitudes averaged within the predifined time windows and over electrodes in the frontal and parietal Regions-
of-Interest, which are shown in the middle. Error bars indicate 95% confidence intervals (ppp,0.01, pppp,0.001). Metaphors evoked an
enhanced frontal negativity relative to Literals both in the N400 (300–500 ms) and the N700 (700–1000 ms) time windows during right visual
field / left hemisphere (LH) presentation. This effect was the more enhanced the more abstract the metaphors were, and it is unique and
specific to the processing of metaphorical senses of concrete words (“sweet”); no such effect has been reported before for novel metaphors
or non-sentential stimuli. Literals evoked a greater frontal negativity in the N700 time window during right hemisphere (RH) relative to LH pre-
sentation, which corresponds to a typical concreteness effect. Both metaphorical and literal novel word pairs evoked a greater parietal neg-
ativity in the N400 time window during LH compared with RH presentation: contrary to some prominent language lateralization models,
novel expressions engaged the LH more than the RH. Grand-average ERP plots over all electrode sites for LH and RH presentation are pro-
vided in Extended Data Figures 1-1,1-2, respectively. Electrophysiological evidence of successful lateralized presentation of target words is
presented in Extended Data Figure 1-3.
Research Article: New Research 5 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
model [n700.ant.lmem = mixed(n700 ;trial 1categoryp-
side 1sidepvalence 1(1 1category|participant) 1(1|item_
nr), n700.ant.data)]. There was a significant main effect of
category (
b
=–0.16, SE= 0.04, F
(1,35.0)
=13.5, p,0.001),
and of valence (
b
=–0.12, SE= 0.04, F
(1,290)
=11.3,
p,0.001), and a category side interaction (
b
=0.14,
SE = 0.04, F
(1,5990)
=11.2, p,0.001). Breaking the interac-
tion down by side showed that metaphors induced a great-
er negativity than Literals during LH presentation (
b
=–
0.29, SE = 0.06, F
(1,34.9)
= 22.2, p,0.001) but not during
RH presentation (
b
=–0.02, SE = 0.06, F
(1,2978)
=0.10,
p= 0.75).
Exploratory analyses of frontal ERPs
It is a question whether literal expressions evoked a
concreteness effect. Huang et al. (2010) reported an en-
hanced frontal negative response for concrete versus ab-
stract word pairs, but only during RH presentation.
Therefore, as a workaround of the lack of an abstract condi-
tion, the significant two-way interaction over the frontal ROI
was broken down also along category (Bonferroni corrected
a
level = 0.0125). Responses evoked by literal expressions
during LH versus RH presentation (Extended Data Fig. 2-2)
were contrasted in the N700 time window, where Lucas et
al. (2017) reported electrophysiological concreteness effects
for novel expressions. Literals indeed evoked a greater neg-
ativity during RH than during LH presentation (
b
=–0.17,
SE = 0.06, F
(1,2913)
= 8.63, p=0.003).
In order to address the potential explanation of the fron-
tal effects based on the first constituents of the word pairs
(as reported by Lucas et al., 2017), noun concreteness
ratings were collected. Nouns were more concrete in liter-
al than in metaphorical expressions (Table 1), yet a greater
frontal negativity was evoked by metaphors relative to lit-
erals, therefore, an explanation of the frontal effects
based on nouns can be excluded.
Metaphors split along median concreteness
To further investigate whether high-concreteness and
low-concreteness metaphors are processed differently
(Forgács et al., 2015), conditions were split along the me-
dian of concreteness (for examples, see Table 2). Both
more abstract and more concrete metaphors were con-
trasted with high-concreteness literals, because these are
the better exemplars of the concrete literal condition and
thus could serve as a stricter baseline with respect to sen-
sorimotor feature processing. Low-concreteness literals
are merely lower on the concreteness scale and are nei-
ther truly abstract nor do they constitute a theoretically
distinct category. Since the adjectives were not identical
in these contrasts, their length and frequency were intro-
duced to the models as a third element of a side cate-
gory interaction (they were included in the final models
only if they significantly improved them). Adjective fre-
quency information was retrieved from the Lexique data-
base (New et al., 2004), and the logarithm of the subtitle-
based word frequency was used, as it has a superior pre-
dictive power compared with other measures (Brysbaert
et al., 2011). Since these latter two variables were of no
theoretical interest and were included only to control for
possible variance, their effect, as well as those of covari-
ates not in interaction with category or side, are reported
in detail only at http://osf.io/j45gt/.
Low-concreteness (abstract) metaphors
First, low-concreteness metaphors were compared
with high-concreteness literals in the 300- to 500-ms
time window. In the final model [n400.ant.Ml.Lh.lmem =
mixed(n400 ;categorypsidepbigram.frequency 1cate-
gorypsidepconcreteness 1(1 1category|participant) 1
(1|item_nr), n400.ant.Ml.Lh.data)], there was a significant
category side concreteness interaction (
b
= 0.29,
SE = 0.14, F
(1,2934)
= 4.60, p= 0.032). When it was broken
Figure 2. Topographical difference maps (metaphor–literal) of ERP responses in the N400 and N700 time windows for each hemi-
field presentation. Blue colors indicate greater negativities for metaphors, relative to literals, which is apparent in both time windows
over frontal areas during LH presentation. Extended Data Figure 2-1 shows topographical maps of the frontal effects in the side
contrast (RH–LH) and demonstrates a typical concreteness for literal expressions in the N700 time window. Extended Data Figures
2-2,2-3 show effects evoked at the prime word (noun).
Research Article: New Research 6 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
down, there was no category main effect during LH pre-
sentation only when concreteness was not included, sug-
gesting that it accounted for the effect. During RH
presentation, a category concreteness interaction was
apparent (
b
= 0.61, SE = 0.19, F
(1,243)
= 9.87, p= 0.002).
When it was further broken down, concreteness affected
ERP responses in opposite directions: high-concreteness
literals followed a typical concreteness effect (
b
=–0.25,
SE = 0.18), albeit non-significant (p= 0.17), but low-con-
creteness metaphors showed a significant inverted effect:
the less concrete (or more abstract) a metaphor was
rated, the greater frontal negativity it evoked (
b
= 0.90,
SE = 0.35, F
(1,90.7)
= 6.49, p= 0.013; Fig. 3).
Next, in the 700- to 1000-ms time window low-con-
creteness metaphors were compared with high-concrete-
ness literals frontally. The final model [n700.ant.Ml.Lh.
lmem = mixed(n700 ;categorypsidepconcreteness 1
sidepvalence 1(1 1category|participant) 1(1 1side|
item_nr), n700.ant.Ml.Lh.data)] showed a significant inter-
action between category concreteness (
b
= 0.47,
SE = 0.15, F
(1,264)
= 9.43, p= 0.002). A category main effect
and a category side interaction was significant only
when Concreteness was not included, which suggests that
it accounted for the effect. When the former interaction was
broken down by category, low-concreteness metaphors
again showed an inverted concreteness effect (or abstract-
ness effect;
b
= 0.84, SE = 0.28, F
(1,94.6)
=9.02, p=0.003),
while high-concreteness literals exhibited a non-significant
(p= 0.60) typical concreteness effect (
b
=–0.08, SE = 0.15;
Fig. 3).
High-concreteness metaphors
High-concreteness metaphors compared with high-con-
creteness literals in the 300- to 500-ms time window [n400.
ant.Mh.Lh.lmem = mixed(n400 ;categorypsidepadjective.
frequency 1sidepvalence 1(1|participant) 1(1|item_nr),
n400.ant.Mh.Lh.data)] revealed a main effect of category
(
b
=–0.12, SE = 0.05, F
(1,1672)
= 5.06, p= 0.025), and of side
(
b
= 0.30, SE = 0.14, F
(1,3026)
= 4.49, p= 0.034), and a tree-
way interaction of category side adjective frequency
interaction (
b
= 0.14, SE = 0.06, F
(1,3029)
=5.29, p= 0.022).
When broken down along side, there was a significant differ-
ence between categories during LH presentation (
b
=
–0.17, SE = 0.07, F
(1,1159)
= 5.72, p= 0.017), but not during
RH presentation (p= 0.39).
Finally, high-concreteness metaphors were contrasted
with high-concreteness literals in the 700- to 1000-ms
time window. The final model [n700.ant.Mh.Lh.lmem =
mixed(n700 ;categorypsidepadjective.frequency 1sidep
valence 1(1|participant) 1(1|item_nr), n700.ant.Mh.Lh.
data)] revealed a main effect of category (
b
=–0.13,
SE = 0.06, F
(1,3088)
= 4.95, p= 0.026), and a three-way inter-
action of category side adjective frequency (
b
=
–0.13, SE = 0.07, F
(1,3098)
=3.98, p= 0.046). When broken
down, there was no effect of condition during RH
(p= 0.83), only LH presentation, where metaphors evoked
a greater negativity (
b
=–0.24, SE = 0.08, F
(1,1512)
=8.93,
p=0.003).
In sum, low-concreteness metaphors, relative to high-
concreteness literals, evoked an abstractness effect dur-
ing RH presentation in the 300- to 500-ms and bilaterally
in the 700- to 1000-ms time windows: the less concrete
(or more abstract) the metaphors were, the greater was
the negativity they evoked. High-concreteness metaphors
induced more negative responses frontally during LH pre-
sentation relative to high-concreteness literals in both
time windows, just like in the main analyses.
Discussion
The present study set out to explore neural responses to
adjectives that refer to concrete, perceptual, physical expe-
riences, when they serve as the figurative part of novel met-
aphorical expressions. The purpose of the experiment was
two-fold: (1) to better understand the role of literal meaning
and sensorimotor feature processing via the electrophysio-
logical concreteness effect when concrete adjectives are
meant in a metaphorical sense; and (2) to test whether the
RHplaysauniqueroleinfigurativelanguageprocessing.
Metaphorical expressions did not evoke a typical cen-
troparietal N400 response relative to literal word pairs.
Table 2: Examples of concrete and abstract, novel metaphorical and literal expressions
Concrete Abstract
tracteur bavard talkative tractor demande brisée broken request
narration mielleuse honeyed narrative affirmation biscornue quirky statement
agenda maigre lean agenda imprudence vide empty recklessness
Metaphor amende piquante zesty fine règlement tricoté knitted settlement
rumeur bouillante boiling rumor oubli dense dense oblivion
pensée enfumée smoky thought phrase graisseuse fat sentence
branche dansante dancing branch concept tuméfié swollen concept
figue croquante crunchy fig batelet usé used dinghy
chouquette délicieuse delicious owl soldat maigriot skinny soldier
mamie hurlante screaming granny catacombe fermée closed catacomb
Literal nectarine piquante zesty nectarine glaive décorée decorated sword
sucrerie pimentée spicy candy conserve empilée stacked can
marmelade coulante flowing marmalade artiste rassasiée satiated artist
lasagne bouillante boiling lasagna colloque vide empty conference
Both conditions were split along their respective median concreteness values. Since less concrete literal expressions are not truly abstract and do not constitute
a theoretically sound category, the best exemplars of the concrete literal condition, high concreteness literals served as baseline for comparison with both high-
concreteness and low-concreteness (i.e., abstract) metaphors.
Research Article: New Research 7 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
Over the frontal ROI literal expressions elicited a electro-
physiological concreteness effect: an enhanced negativity
in the 700- to 1000-ms time window during RH relative
to LH presentation. The least negativity was evoked by
literals in the LH, where metaphors, compared to literals,
evoked an enhanced frontal negativity in both the 300-
to 500-ms and the 700- to 1000-ms time windows, a
kind of effect that previously has not been reported with
non-sentential stimuli. During RH presentation meta-
phors evoked the same level of negative amplitudes in
the 300- to 500-ms time window as literals, however,
these were the more negative the less concrete (or more
abstract) the metaphors were, just the opposite of the
typical concreteness effect. The frontal effect meta-
phors evoked in the 700- to 1000-ms time window was
more negative bilaterally for increasingly more abstract
metaphors. Neural responses to metaphors were not
driven by figurativeness or imageability, but abstract-
ness. RH presentation did not elicit a distinct response
to metaphors, and the findings do not support an
embodied account of semantic processing either, since sen-
sorimotor feature processing, as reflected by a typical con-
creteness effects, was observed only for literal language.
The first important result is the lack of a typical centro-
parietal N400 effect for metaphors compared with literal
expressions. Several studies reported an enhanced N400
effect when participants read novel metaphors, and most
studies interpreted it as reflecting conceptual mappings
and/or blends (Coulson and Van Petten, 2002,2007;
Arzouan et al., 2007;Lai et al., 2009;Goldstein et al.,
2012;Rutter et al., 2012;Lai and Curran, 2013;Schneider
et al., 2014;Tang et al., 2017;Rataj et al., 2018) or as the
activation of literal senses (Pynte et al., 1996;Tartter et
al., 2002;De Grauwe et al., 2010;Weiland et al., 2014).
Such findings could have been artifacts of stimulus de-
signs that did not take into consideration the novelty of
metaphorical and literal control conditions. Metaphors
might not elicit a typical N400 (Yang et al., 2013;Bardolph
and Coulson, 2014) and might not be processed via
blends, mappings, or literal senses.
Figure 3. Abstractness effect evoked by metaphors over the frontal ROI. Upper panel shows ERP responses for individual trials to
low-concreteness metaphors and high-concreteness literals. Electrophysiological responses over frontal electrode sites were the
more negative, the less concrete (i.e., more abstract) the metaphors were both in in the 300- to 500-ms time window during RH pre-
sentation and also in the 700- to 1000-ms time window bilaterally. Lower panel shows bar charts of the median split high-concrete-
ness and low-concreteness metaphors relative to median split high-concreteness literals. Stars in brackets (p) indicate effects that
were significant if concreteness was not included in the models (i.e., concreteness accounted for these effects). Error bars indicate
95% confidence intervals (pp,0.05, ppp,0.01).
Research Article: New Research 8 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
The most remarkable outcome is a frontal response to
metaphors both in the N400 and the N700 time windows
during LH presentation. It conforms with a number of find-
ings reporting LH effects for novel metaphors (Rapp et al.,
2004,2007;Diaz and Hogstrom, 2011;Forgács et al.,
2012,2014;Rutter et al., 2012). However, only Coulson
and Van Petten (2007) reported specifically late left frontal
negativities for metaphors, which they explained as a re-
duced frontal positivity because of low information selec-
tion demands. Instead of interpreting the effect as a
reduced electrophysiological positivity, the frontal nega-
tivity for metaphors reported here could be an ERP re-
sponse on its own right.
Its neural generators might be similar to or partially
overlapping with those underlying left anterior negativities
(LANs), which have been typically reported with sentential
stimuli. It has been proposed that in the 300- to 450-ms
time window LAN is sensitive to various morphosyntactic
operations (Molinaro et al., 2015), for example, the viola-
tion of syntactic expectancy (Molinaro et al., 2011). A sus-
tained LAN has been attributed to syntactic working
memory load preceding syntactic integration (Fiebach et
al., 2002), which can be evoked by function words (Neville
et al., 1992;Kluender and Kutas, 1993) but also during
thematic role assignment (King and Kutas, 1995). In their
review, Federmeier and Laszlo (2009) concluded that a
post-N400 frontal negativity could be related to meaning
selection during ambiguity resolution. Novel metaphors
plausibly require selecting an appropriate figurative sense
of concrete adjectives, finding a link with the noun, and in-
tegrating the two into a unified representation. Developing
the relational structure of the constituents and construing
the properties of a combined representation could require
syntactic-like operations and could easily pose working
memory demands.
The frontal effects are clearly distinct from the electro-
physiological concreteness effect. Lucas et al. (2017)
found an enhanced frontal negativity in the N400 time
window for abstract word pairs, which they interpreted as
a positivity for concrete words pairs, driven by responses
to first constituents. However, their results do not provide
evidence for the additive nature of these effects, and an
analysis of ERPs at the noun in the current study revealed
no frontal concreteness effect, which could have ex-
plained later outcomes (Extended Data Figs. 2-2,2-3).
The lack of a difference between the metaphorical and
literal conditions during RH presentation could imply
an equal amount of sensorimotor feature processing.
However, concreteness influenced amplitudes in the op-
posite direction for metaphors and literals. Literals evoked
a typical concreteness effect, but metaphors evoked the
greater frontal negativities the more abstract they were.
This outcome replicates Forgács et al. (2015)’s result that
abstract metaphors evoked a stronger N400 effect, not
concrete ones. The ideas that novel metaphors rely on lit-
eral meanings (Bowdle and Gentner, 2005), that they are
based on concrete source domains, involve mappings
(Lakoff, 2014), or evoke sensorimotor processes (Gallese
and Lakoff, 2005), beyond early automatic activations,
which are not contested, did not receive empirical
support. If metaphors do not transmit sensorimotor infor-
mation via mappings to abstract target domains, and em-
bodiment concerns only concrete literal language, its
explanatory power for the conceptual system and cogni-
tion in general might be rendered rather limited.
Kousta et al. (2011) reported an abstractness effect in
terms of reaction times, which challenged decades of
concreteness effect research. The neural pattern of the
electrophysiological abstractness effect contradicts
Paivio (2007)’s dual coding theory, which suggests that all
words activate a purely linguistic, amodal code in the LH,
while concrete words activate an additional imagistic
code in the RH. First, the concrete adjectives did not elicit
a concreteness effect when used figuratively. Second, the
abstractness effect was bilateral in the N700 time win-
dow, which questions the lateralized implementation of
the two codes. Third, the fact that the RH produced a
concreteness effect for literals in the N700 time window
(replicating Huang et al., 2010;Lucas et al., 2017), while it
contributed to an abstractness effect for metaphors in
both time windows suggests that there is at least partial
overlap between the neural generators of the verbal and
imagistic codes, they are not lateralized in the two hemi-
spheres. Further studies are necessary to specify the na-
ture of the electrophysiological abstractness effect and its
relation to sensorimotor feature processing.
One more notable finding of the present study is the left
lateralization of the typical N400 both for novel metaphorical
and literal stimuli (replicating Diaz and Hogstrom, 2011;
Forgács et al., 2012,2014;Davenport and Coulson, 2013).
Both hemispheres are able to produce an N400 (Federmeier
et al., 2005), which is indicative of a greater semantic memo-
ry retrieval effort (Kutas and Federmeier, 2011)ormeaning
activation (Molinaro et al., 2010). Prominent language later-
alization models, such as the coarse semantic coding theory
(Jung-Beeman, 2005) and the graded salience hypothesis
(Giora, 2003), predict a greater RH activation during the
comprehension of novel linguistic constructions. No such
outcome was observed. Van Lancker Sidtis (2004)’sdual-
process model can account for the current LH results: novel
language, regardless of figurativeness, shall tax the LH be-
cause of effortful meaning construction. A curious finding is
that emotional valence, which has been reported to evoke
stronger responses when it is more negative (Altmann et al.,
2012), did so in the current study when it was more positive,
reflecting perhaps the joy of reading unfamiliar, creative
expressions.
What processing steps could the electrophysiological
response pattern indicate? The typical N400 response
suggests a greater semantic activation (Molinaro et al.,
2010;Kutas and Federmeier, 2011) in the LH for all novel
word pairs. It is accompanied by an enhanced frontal
N400 component in the LH for metaphors, which poten-
tially reflects some sort of rule-based process (Molinaro et
al., 2011,2015), perhaps morphosyntactic/structural/con-
ceptual combinatorics. This greater frontal negativity for
metaphors is sustained during the 700- to 1000-ms time
window, which could be related to (syntactic) working
memory load (Fiebach et al., 2002), and/or ambiguity re-
solution and meaning selection (Federmeier and Laszlo,
Research Article: New Research 9 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
2009). The frontal metaphor effects were sensitive to ab-
stractness in the RH during the N400 time window and bi-
laterally during the N700 time window, which could be
analogous to the conceptual manipulations proposed for
the typical concreteness effect: the N400 for the activa-
tion (Gullick et al., 2013;Lucas et al., 2017) and the N700
for the conceptual integration (Barber et al., 2013), but in
this case, of abstract properties. The suppression and en-
hancement of properties, which have been reported for
metaphors (Gernsbacher et al., 2001) and for ambiguous
words (Gergely and Pléh, 1994), could be reflected in
these ERP responses. The adjective’s concrete, literal
senses need to be suppressed and an appropriate, ab-
stract, figurative sense needs to be activated, selected,
and integrated into the representation of the expression.
The data are not inconsistent with Glucksberg (2003)’s
categorization account, however, instead of describing
the non-perceptual senses of metaphors in terms of
superordinate categories and characterizing property se-
lection along literal meaning (Glucksberg et al., 2001), ab-
stractness might hold the key to figurative meaning. The
abstract conceptual substitution model of metaphor
comprehension (Forgács, 2014) proposes that the en-
hancement and suppression of conceptual properties
(Gernsbacher et al., 2001) is carried out along the ab-
stract-concrete dimension: not basic-level (or “literal”), but
concrete properties are suppressed, and not superordinate
(or “figurative”) but abstract properties are enhanced. From
this narrower set of abstract senses, during the construction
of metaphorical meaning, the contextually most relevant ab-
stract property is substituted in the place of the figuratively
meant word. The model does not require the construction of
ad hoc categories (Glucksberg, 2003) or ad hoc concepts
(Carston, 2010), not even metaphorical mappings (Lakoff
and Johnson, 1980) or conceptual blends (Fauconnier and
Turner, 1998), but suggests a unique form of ambiguity re-
solution for metaphors. Meaning filtering processes, similar
to those employed for polysemous words (Murphy, 1997),
could be used along abstractness, as a means of figurative
“meaning making”(Bruner, 1990), a creative process of es-
tablishing, and not looking up, an interpretation. The theory
moves beyond the literal-figurative distinction and can ac-
count for the reported abstractness effect as well: the more
abstract the overall sense of the figurative expression is, the
more effort could the search for the most appropriate ab-
stract property require.
The present study reports of a frontal, negative going
brain wave evoked by novel metaphors during LH presen-
tation, which is sensitive to their abstractness, regardless
of their meaningfulness, imageability, or figurativeness.
The experiment did not find evidence for the RH theory of
metaphor, for the involvement of the literal meaning of
novel metaphors (Bowdle and Gentner, 2005), for the
strong version of embodiment (Gallese and Lakoff, 2005;
Lakoff, 2014), for some claims of Paivio (2007)’s dual cod-
ing theory, and for some prominent language lateraliza-
tion models either (Giora, 2003;Jung-Beeman, 2005).
Novel language appears to be processed by the LH and
formulaic language by the RH (Van Lancker Sidtis, 2004),
and metaphor appears to be no exception. Based on the
data reported here, a novel picture of metaphor compre-
hension seems to emerge. The processing of figurative
senses might not depend on concrete, sensorimotor fea-
tures but on the semantic manipulation of abstract prop-
erties of words used metaphorically.
References
Altmann U, Bohrn IC, Lubrich O, Menninghaus W, Jacobs AM (2012)
The power of emotional valence—from cognitive to affective proc-
esses in reading. Front Hum Neurosci 6:192.
Arzouan Y, Goldstein A, Faust M (2007) Brainwaves are stetho-
scopes: ERP correlates of novel metaphor comprehension. Brain
Res 1160:69–81.
Baayen RH (2008) Analyzing linguistic data: a practical introduction
to statistics using R. Cambridge: Cambridge University Press.
Baayen RH, Davidson DJ, Bates DM (2008) Mixed-effects modeling
with crossed random effects for subjects and items. J Mem Lang
59:390–412.
Barber HA, Otten LJ, Kousta ST, Vigliocco G (2013) Concreteness in
word processing: ERP and behavioral effects in a lexical decision
task. Brain Lang 125:47–53.
Bardolph M, Coulson S (2014) How vertical hand movements impact
brain activity elicited by literally and metaphorically related words:
an ERP study of embodied metaphor. Front Hum Neurosci 8:1031.
Barr DJ, Levy R, Scheepers C, Tily HJ (2013) Random effects struc-
ture for confirmatory hypothesis testing: keep it maximal. J Mem
Lang 68:255–278.
Bates D, Mächler M, Bolker B, Walker S (2015) Fitting linear mixed-
effects models using lme4. J Stat Soft 67:1–48.
Binder JR, Desai RH (2011) The neurobiology of semantic memory.
Trends Cogn Sci 15:527–536.
Bohrn IC, Altmann U, Jacobs AM (2012) Looking at the brains behind
figurative language –a quantitative meta-analysis of neuroimaging
studies on metaphor, idiom, and irony processing. Neuropsychologia
50:2669–2683.
Boulenger V, Hauk O, Pulvermüller F (2009) Grasping ideas with the
motor system: semantic somatotopy in idiom comprehension.
Cereb Cortex 19:1905–1914.
Boulenger V, Shtyrov Y, Pulvermüller F (2012) When do you grasp
the idea? MEG evidence for instantaneous idiom understanding.
Neuroimage 59:3502–3513.
Bowdle B, Gentner D (2005) The career of metaphor. Psychol Rev
112:193–216.
Brouwer H, Fitz H, Hoeks J (2012) Getting real about semantic illu-
sions: rethinking the functional role of the P600 in language com-
prehension. Brain Res 1446:127–143.
Bruner JS (1990) Acts of meaning. Cambridge: Harvard University
Press.
Brysbaert M, Buchmeier M, Conrad M, Jacobs AM, Bölte J, Böhl A
(2011) The word frequency effect. Exp Psychol 58:412–424.
Cardillo ER, Watson CE, Schmidt GL, Kranjec A, Chatterjee A (2012)
From novel to familiar: tuning the brain for metaphors. Neuroimage
59:3212–3221.
Carston R (2010) XIII—metaphor: ad hoc concepts, literal meaning
and mental images. Proc Aristot Soc 110:295–321.
Chen E, Widick P, Chatterjee A (2008) Functional–anatomical organi-
zation of predicate metaphor processing. Brain Lang 107:194–
202.
Citron FM, Goldberg AE (2014) Metaphorical sentences are more
emotionally engaging than their literal counterparts. J Cogn
Neurosci 26:2585–2595.
Coulson S, Van Petten C (2002) Conceptual integration and meta-
phor: an event-related potential study. Mem Cognit 30:958–968.
Coulson S, Van Petten C (2007) A special role for the right hemi-
sphere in metaphor comprehension? ERP evidence from hemifield
presentation. Brain Res 1146:128–145.
Research Article: New Research 10 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
Davenport T, Coulson S (2013) Hemispheric asymmetry in interpret-
ing novel literal language: an event-related potential study.
Neuropsychologia 51:907–921.
De Grauwe S, Swain A, Holcomb PJ, Ditman T, Kuperberg GR (2010)
Electrophysiological insights into the processing of nominal meta-
phors. Neuropsychologia 48:1965–1984.
Desai RH, Binder JR, Conant LL, Mano QR, Seidenberg MS (2011)
The neural career of sensory-motor metaphors. J Cogn Neurosci
23:2376–2386.
Diaz MT, Hogstrom LJ (2011) The influence of context on hemi-
spheric recruitment during metaphor processing. J Cogn Neurosci
23:3586–3597.
Fauconnier G, Turner M (1998) Conceptual integration networks.
Cognitive Science 22:133–187.
Federmeier KD, Laszlo S (2009) Time for meaning: electrophysiology
provides insights into the dynamics of representation and proc-
essing in semantic memory. Psychol Learn Motiv 51:1–44.
Federmeier KD, Mai H, Kutas M (2005) Both sides get the point: hem-
ispheric sensitivities to sentential constraint. Mem Cognit 33:871–
886.
Fiebach CJ, Schlesewsky M, Friederici AD (2002) Separating syntac-
tic memory costs and syntactic integration costs during parsing:
the processing of German WH-questions. J Mem Lang 47:250–
272.
Forgács B (2014) Figures of language in cognitive science in the light
of figurative language processing in the brain (Unpublished doc-
toral dissertation). Budapest: Budapest University of Technology
and Economics. Available from http://hdl.handle.net/10890/1336.
Forgács B, Bohrn I, Baudewig J, Hofmann MJ, Pléh C, Jacobs AM
(2012) Neural correlates of combinatorial semantic processing of
literal and figurative noun noun compound words. Neuroimage
63:1432–1442.
Forgács B, Lukács Á, Pléh C (2014) Lateralized processing of novel meta-
phors: disentangling figurativeness and novelty. Neuropsychologia
56:101–109.
Forgács B, Bardolph M, Amsel BD, DeLong KA, Kutas M (2015)
Metaphors are physical and abstract: ERPs to metaphorically
modified nouns resemble ERPs to abstract language. Front Hum
Neurosci 9:28.
Gallese V, Lakoff G (2005) The brain’s concepts: the role of the sen-
sory-motor system in conceptual knowledge. Cogn Neuropsychol
22:455–479.
Gergely G, Pléh C (1994) Lexical processing in an agglutinative lan-
guage and the organization of the lexicon. Folia Linguistica
28:175–204.
Gernsbacher MA, Keysar B, Robertson RR, Werner NK (2001) The
role of suppression and enhancement in understanding meta-
phors. J Mem Lang 45:433–450.
Giora R (2003) On our mind: salience, context and figurative lan-
guage. New York: Oxford University Press.
Glucksberg S (2003) The psycholinguistics of metaphor. Trends
Cogn Sci 7:92–96.
Glucksberg S, Newsome MR, Goldvarg Y (2001) Inhibition of the lit-
eral: filtering metaphor-irrelevant information during metaphor
comprehension. Metaph Symbol 16:277–298.
Goldstein A, Arzouan Y, Faust M (2012) Killing a novel metaphor and
reviving a dead one: ERP correlates of metaphor conventionaliza-
tion. Brain Lang 123:137–142.
Gullick MM, Mitra P, Coch D (2013) Imagining the truth and the
moon: an electrophysiological study of abstract and concrete
word processing. Psychophysiology 50:431–440.
Holcomb PJ, Kounios J, Anderson JE, West WC (1999) Dual-coding,
context-availability, and concreteness effects in sentence compre-
hension: an electrophysiological investigation. J Exp Psychol
Learn Mem Cogn 25:721–742.
Huang HW, Lee CL, Federmeier KD (2010) Imagine that! ERPs pro-
vide evidence for distinct hemispheric contributions to the proc-
essing of concrete and abstract concepts. Neuroimage 49:1116–
1123.
Jung-Beeman M (2005) Bilateral brain processes for comprehending
natural language. Trends Cogn Sci 9:512–518.
King JW, Kutas M (1995) Who did what and when? Using word-and
clause-level ERPs to monitor working memory usage in reading. J
Cogn Neurosci 7:376–395.
Kluender R, Kutas M (1993) Bridging the gap: evidence from ERPs
on the processing of unbounded dependencies. J Cogn Neurosci
5:196–214.
Kounios J, Holcomb PJ (1994) Concreteness effects in semantic
processing: ERP evidence supporting dual-coding theory. J Exp
Psychol Learn Mem Cogn 20:804–823.
Kousta ST, Vigliocco G, Vinson DP, Andrews M, Del Campo E (2011)
The representation of abstract words: why emotion matters. J Exp
Psychol 140:14–34.
Kutas M, Federmeier KD (2011) Thirty years and counting: finding
meaning in the N400 component of the event-related brain poten-
tial (ERP). Annu Rev Psychol 62:621–647.
Lacey S, Stilla R, Sathian K (2012) Metaphorically feeling: compre-
hending textural metaphors activates somatosensory cortex.
Brain Lang 120:416–421.
Lacey S, Stilla R, Deshpande G, Zhao S, Stephens C, McCormick K,
Kemmerer D, Sathian K (2017) Engagement of the left extrastriate
body area during body-part metaphor comprehension. Brain Lang
166:1–18.
Lai VT, Curran T (2013) ERP evidence for conceptual mappings and
comparison processes during the comprehension of conventional
and novel metaphors. Brain Lang 127:484–496.
Lai VT, Curran T, Menn L (2009) Comprehending conventional and
novel metaphors: an ERP study. Brain Res 1284:145–155.
Lakoff G (2014) Mapping the brain’s metaphor circuitry: metaphorical
thought in everyday reason. Front Hum Neurosci 8:958.
Lakoff G, Johnson M (1980) Metaphors we live by, Ed 2. Chicago:
University of Chicago Press.
Lakoff G, Johnson M (1999) Philosophy in the flesh: the embodied
mind and its challenge to Western thought. New York: Basic
Books.
Lee CL, Federmeier KD (2008) To watch, to see, and to differ: an
event-related potential study of concreteness effects as a function
of word class and lexical ambiguity. Brain Lang 104:145–158.
Lucas HD, Hubbard RJ, Federmeier KD (2017) Flexible conceptual
combination: electrophysiological correlates and consequences
for associative memory. Psychophysiology 54:833–847.
Mahon BZ, Caramazza A (2008) A critical look at the embodied cog-
nition hypothesis and a new proposal for grounding conceptual
content. J Physiol 102:59–70.
Mithen SJ (1996) The prehistory of the mind: a search for the origins
of art, religion, and science. London: Thames and Hudson.
Molinaro N, Conrad M, Barber HA, Carreiras M (2010) On the func-
tional nature of the N400: contrasting effects related to visual word
recognition and contextual semantic integration. Cogn Neurosci
1:1–7.
Molinaro N, Barber HA, Carreiras M (2011) Grammatical agreement
processing in reading: ERP findings and future directions. Cortex
47:908–930.
Molinaro N, Barber HA, Caffarra S, Carreiras M (2015) On the left an-
terior negativity (LAN): the case of morphosyntactic agreement.
Cortex 66:156–159.
Murphy GL (1997) Reasons to doubt the present evidence for meta-
phoric representation. Cognition 62:99–108.
Neville HJ, Mills DL, Lawson DS (1992) Fractionating language: dif-
ferent neural subsystems with different sensitive periods. Cereb
Cortex 2:244–258.
New B, Pallier C, Brysbaert M, Ferrand L (2004) Lexique 2: a new
French lexical database. Behav Res Methods Instrum Comput
36:516–524.
Oldfield RC (1971) The assessment and analysis of handedness: the
Edinburgh inventory. Neuropsychologia 9:97–113.
Paivio A (2007) Mind and its evolution: a dual coding theoretical ap-
proach. Mahwah: L. Erlbaum Associates.
Research Article: New Research 11 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
Pinker S (2010) The cognitive niche: coevolution of intelligence, so-
ciality, and language. Proc Natl Acad Sci USA 107:8993–8999.
Pomp J, Bestgen AK, Schulze P, Müller CJ, Citron FM, Suchan B,
Kuchinke L (2018) Lexical olfaction recruits olfactory orbitofrontal
cortex in metaphorical and literal contexts. Brain Lang 179:11–21.
Pulvermüller F (2005) Brain mechanisms linking language and action.
Nat Rev Neurosci 6:576–582.
Pulvermüller F, Shtyrov Y, Ilmoniemi R (2005) Brain signatures of
meaning access in action word recognition. J Cogn Neurosci
17:884–892.
Pynte J, Besson M, Robichon F-H, Poli J (1996) The time-course of
metaphor comprehension: an event-related potential study. Brain
Lang 55:293–316.
Core Team R (2017) R: a language and environment for statistical
computing. Vienna: R Foundation for Statistical Computing.
Rapp AM, Leube DT, Erb M, Grodd W, Kircher TTJ (2004) Neural cor-
relates of metaphor processing. Cogn Brain Res 20:395–402.
Rapp AM, Leube DT, Erb M, Grodd W, Kircher TTJ (2007) Laterality
in metaphor processing: lack of evidence from functional magnetic
resonance imaging for the right hemisphere theory. Brain Lang
100:142–149.
Rataj K, Przekoracka-Krawczyk A, van der Lubbe RH (2018) On
understanding creative language: the late positive complex and
novel metaphor comprehension. Brain Res 1678:231–244.
Rutter B, Kröger S, Hill H, Windmann S, Hermann C, Abraham A
(2012) Can clouds dance? Part 2: an ERP investigation of passive
conceptual expansion. Brain Cogn 80:301–310.
Saygin AP, McCullough S, Alac M, Emmorey K (2010) Modulation of
BOLD response in motion-sensitive lateral temporal cortex by real
and fictive motion sentences. J Cogn Neurosci 22:2480–2490.
Schmidt-Snoek GL, Drew AR, Barile EC, Agauas SJ (2015) Auditory
and motion metaphors have different scalp distributions: an ERP
study. Front Hum Neurosci 9:126.
Schneider S, Rapp AM, Haeußinger FB, Ernst LH, Hamm F,
Fallgatter AJ, Ehlis AC (2014) Beyond the N400: complementary
access to early neural correlates of novel metaphor comprehen-
sion using combined electrophysiological and haemodynamic
measurements. Cortex 53:45–59.
Seidenberg MS, Tanenhaus MK, Leiman JM, Bienkowski M (1982)
Automatic access of the meanings of ambiguous words in context:
some limitations of knowledge-based processing. Cogn Psychol
14:489–537.
Shen ZY, Tsai YT, Lee CL (2015) Joint influence of metaphor familiar-
ity and mental imagery ability on action metaphor comprehension:
An event-related potential study. Language and Linguistics
16:615–637.
Shtyrov Y, Butorina A, Nikolaeva A, Stroganova T (2014) Automatic
ultrarapid activation and inhibition of cortical motor systems in
spoken word comprehension. Proc Natl Acad Sci USA 111:
E1918–E1923.
Singmann H, Kellen D (2017) An introduction to mixed models for ex-
perimental psychology. In: New methods in neuroscience and
cognitive psychology (Spieler DH and SchumacherE, eds).
London: Routledge.
Swinney D (1979) Lexical access during sentence comprehension:
(re)consideration of context effects. J Verbal Learn Verbal Behav
18:645–659.
Tang X, Qi S, Jia X, Wang B, Ren W (2017) Comprehension of scien-
tific metaphors: complementary processes revealed by ERP. J
Neurolinguistics 42:12–22.
Tartter VC, Gomes H, Dubrovsky B, Molholm S, Stewart RV (2002)
Novel metaphors appear anomalous at least momentarily: evi-
dence from N400. Brain Lang 80:488–509.
Van Lancker Sidtis D (2004) When novel sentences spoken or heard
for the first time in the history of the universe are not enough: to-
ward a dual-process model of language. Int J Lang Commun
Disord 39:1–44.
Weiland H, Bambini V, Schumacher PB (2014) The role of literal
meaning in figurative language comprehension: evidence from
masked priming ERP. Front Hum Neurosci 8:583.
West WC, Holcomb PJ (2000) Imaginal, semantic, and surface-level
processing of concrete and abstract words: an electrophysiologi-
cal investigation. J Cogn Neurosci 12:1024–1037.
Yang FPG, Bradley K, Huq M, Wu DL, Krawczyk DC (2013)
Contextual effects on conceptual blending in metaphors: an
event-related potential study. J Neurolinguistics 26:312–326.
Research Article: New Research 12 of 12
September/October 2020, 7(5) ENEURO.0052-20.2020 eNeuro.org
