Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Neuroimage. Author manuscript; available in PMC 2011 January 1.
Published in final edited form as:
PMCID: PMC2782386

Imagine that! ERPs provide evidence for distinct hemispheric contributions to the processing of concrete and abstract concepts


Although abstract and concrete concepts are processed and remembered differently, the underlying nature of those differences remains in dispute. The current study used visual half-field (VF) presentation methods and event-related potential (ERP) measures to examine how the left (LH) and right (RH) cerebral hemispheres process concrete and abstract meanings of polysemous nouns (e.g., “green book,” referring to the concrete, physical object that is a book, versus “engaging book,” referring to the abstract information that a book conveys). With presentation to the right VF, nouns preceded by concrete modifiers were associated with more positivity on the P2 and N400, suggesting that concrete concepts were easier for the LH to process perceptually and semantically. In contrast, with presentation to the left VF (RH), nouns used in a concrete sense elicited a sustained frontal negativity (500-900 ms) that has been previously linked to imagery. The results thus reveal multiple, distinct neural and cognitive sources for concreteness effects and point to a critical role for the RH in linking language input to sensory imagery.

In the prologue to Shakespeare's Hamlet, the chorus encourages the audience to:

Think when we talk of horses, that you see them

Printing their proud hoofs i' the receiving earth;

For 'tis your thoughts that now must deck our kings.

The power of language to evoke vivid sensory imagery remembered or newly created has long been celebrated by authors, harnessed by advertisers and politicians, and recognized by language users of all types. Research over several decades has shown that the processing of concrete expressions – i.e., those that denote a concept that can easily be experienced by the senses – is facilitated relative to that for more abstract expressions in a wide range of language, memory, and other higher cognitive tasks. Concrete words are easier to identify (Schwanenflugel et al, 1988), read aloud (Gerhand and Barry, 2000) and remember (Paivio, 1991). Similarly, concrete sentences are easier to comprehend (Schwanenflugel and Shoben, 1983) and to judge for meaningfulness and truthfulness (Belmore et al., 1982). Yet the underlying nature of the processing differences between concrete and abstract expressions remains poorly understood.

One view, known as “context availability theory” (Schwanenflugel, 1991), hypothesizes that concreteness effects arise through differences in the extent and strength of the associative connections between these word types and information stored in a single, amodal semantic system (e.g., Pylyshyn, 1984). On this account, then, concreteness effects reflect quantitative processing differences within a single neurocognitive system, which should be reduced or eliminated by contextual information that provides more semantic support for abstract expressions.

Studies examining the basis for concreteness effects using visual half-field (VF) presentation (to induce asymmetric processing biases) or hemodynamic brain imaging methods, however, have suggested instead that concrete and abstract expressions may be processed by at least partially non-overlapping neural systems distributed across the two cerebral hemispheres (e.g., Binder et al., 2005; Eviatar et al., 1990; Sabsevitz et al., 2005). These patterns provide support for views emphasizing that language comprehension involves the simulation of embodied experience, including modality-specific perceptual and motor processes (e.g., Glenberg and Robertson, 2000; Lakoff & Johnson, 1999).

For example, Paivio's (2007) “dual coding theory,” says that concrete words manifest processing advantages because they are encoded both verbally, in a left-hemisphere (LH) language system, and nonverbally, using sensory imagery that arises from bilateral processing mechanisms. Although many studies provide support for the general hypothesis that the left and right (RH) cerebral hemispheres differ in their processing of concrete and abstract words, the precise pattern of laterality is less clear. Some studies have reported patterns that match the predictions of dual coding, showing bilateral processing of concrete words but LH-dominant processing of abstract words (Binder et al., 2005; Eviatar et al., 1990; Sabsevitz et al., 2005). Other studies, however, observed increased LH activity (Fiebach and Friederici, 2003) and LH processing advantages (see review in Chiarello et al., 2001) for concrete words, or enhanced activations for abstract words bilaterally (Pexman et al. 2007) or selective to the LH (Noppeney and Price, 2004) or RH (Kiehl et al., 1999).

The disparity across studies suggests that concreteness effects may involve multiple mechanisms, which are differentially engaged by different types of stimuli and/or under different task conditions (e.g., Chiarello et al., 2001). This view receives additional support from event-related potential (ERP) studies. Holcomb and colleagues (Holcomb et al., 1999; Kounios and Holcomb, 1994; West and Holcomb, 2000) have shown that concrete words elicit more negative-going potentials beginning in the time window of the N400, a component that has been closely linked to semantic processing (Federmeier and Laszlo, in press). Consistent with aspects of context availability theory, these differences are attenuated when concrete and abstract words are put into predictive sentence contexts (Holcomb et al., 1999) or are repeated (Kounios and Holcomb, 1994). However, whereas the typical N400 has a centro-posterior scalp distribution, the potentials elicited by concrete words between 250 and 500 ms are also more negative over the front of the head, and this frontal negative effect is sustained beyond the N400 time window. These distributional and timecourse differences suggest non-identical neural sources involved in the processing of concrete and abstract words, in line with dual coding.

There are other indications that the electrophysiological pattern reflects dissociable mechanisms. The more posterior (N400) part of the concreteness effect is consistent across word types (appearing similarly for nouns, verbs, and class-ambiguous words; Lee and Federmeier, 2008) and, consistent with the larger literature on N400 effects (for a review, see Kutas et al., 2007), is sensitive to context (Holcomb et al., 1999; Swaab et al., 2002). In contrast, the sustained frontal effect is sensitive to word type (appearing for nouns and unambiguous, but not ambiguous, verbs; Lee and Federmeier, 2008) and task (larger for semantic and imagery tasks than for lexical decision; West and Holcomb, 2000), but less sensitive to the availability of context information (Swaab et al., 2002). Frontally-distributed N400 potentials are also seen to pictures (e.g., Ganis et al, 1996) and effects with a similar distribution (taking into account differences in reference site) have been observed in ERP studies of mental imagery (Farah et al., 1989).

If the frontal concreteness effect is due to mental imagery, then both dual coding theory (Paivio, 2007) and the more general literature on imagery (e.g., Ehrlichman and Barrett, 1983; Kosslyn, 1987) would posit links between this effect and right hemisphere (RH) processing. Thus, a primary goal of the present study is to further elucidate the neural mechanisms involved in the processing of concrete expressions by combining ERP measures with VF methods in order to look at hemispheric processing biases. We predict important RH contributions to the frontal concreteness effect, pointing to a unique role for the RH in mediating the access of perceptually rich meaning information during language comprehension.

At the same time, we further examine the N400 concreteness effect, which seems likely to arise from general mechanisms of semantic access (Federmeier & Laszlo, in press). A difficulty with interpreting the extant data is that effects have been assessed across different sets of lexical items, which, in addition to differing in concreteness, potentially differ along other dimensions known to affect N400 amplitude (e.g., orthographic neighborhood size and semantic ambiguity). In the present study, therefore, we adopt the novel approach of manipulating concreteness on the same lexical items by taking advantage of polysemy. We use (lateralized) nouns that have both a concrete and an abstract meaning (e.g., “book”, which can refer to a physical object or to the information conveyed by that object) and highlight one of these senses through modification by a centrally-presented adjective – for example, “green book” versus “engaging book”. Whereas the RH may be more involved in sensory imagery, as indexed by frontally-distributed potentials, we predict that the LH may be more likely to use information provided by the adjective context to adjust its linguistic (N400) and even perceptual (P2) processing of the polysemous noun (as suggested by the PARLO hypothesis; Federmeier, 2007).


Thirty-two University of Illinois undergraduates (16 females, mean age 20, range 18-25) participated for cash or course credit. All were monolingual English speakers with no early second language exposure and no history of neurological or psychiatric disorders. Participants were right-handed, as determined by the Edinburgh inventory (Oldfield, 1971); 12 reported left-handed family members.

Stimuli consisted of 112 nouns with multiple, related meaning senses (i.e., polysemous nouns), which varied in their baseline concreteness (Toglia and Battig, 1978) from fairly abstract (e.g., “atmosphere” and “vision”) to concrete (“nose” and “lamb”). Adjectives were then used to pick out concrete and abstract senses of each of these polysemous nouns, making use of well-studied polysemous alternations such as product/producer, building/institution, and animal/meat, among others (e.g., Pustejovsky, 1995). An adjective referring to physical aspects of the noun (e.g., “hilly farm”) was used to induce a concrete reading, whereas an adjective referring to features less available to the senses (e.g., “productive farm”) was used to induce an abstract reading. Each noun was paired with four adjectives, two concrete and two abstract, resulting in a final set of 448 adjective-noun pairs.

The perceived concreteness of the adjective-noun pairs was assessed in a norming study. Forty-four monolingual English-speaking University of Illinois students (19 females, mean age 19, range 18-25) participated for course credit. No participant was run in both the norming and ERP studies. Word pairs were divided into two lists; within each, nouns were repeated once, paired with different adjectives and separated by at least 50 pairs. Participants used a 1-7 concreteness rating scale, with “1” indicating a concept that “is difficult to experience with one's senses” (i.e., is abstract). The final set of concrete adjective-noun pairs had a mean rating of 6.01 (SE = .04) and the final set of abstract pairs had a mean rating of 2.96 (SE = .06).

Critical items were divided into two sets, containing 112 each of concrete and abstract adjective-noun pairs followed by probe adjectives (drawn from the same normed set, half concrete and half abstract), which were used for the modification appropriateness judgment task. Within a set, nouns were repeated once across VF (at least 100 trials apart), appearing once in a concrete and once in an abstract pairing. Two additional lists, identical except that presentation VF was reversed, were created from each set. These four lists were then used to create another four in which the critical and probe adjectives were swapped. To each of the lists, 168 filler triplets were added, in which one (56 items) or both (112 items) of the adjectives were anomalous as a modifier of the noun; two-thirds of these nouns were repeated across VF, always with different adjectives. Stimuli were randomized once for each list and presented to participants in the same order. In total, each participant saw 392 triplets: 56 concrete pairs, 56 abstract pairs and 84 filler trials in each VF. Table 1 shows an example of the stimuli and list structure.

Table 1
Example stimuli and list structure

Participants viewed the stimuli sitting one meter in front of a computer screen in a dimly lit room. Their task was to read each three-word series (adjective, lateralized noun, adjective) without moving their eyes from central fixation (a 3 by 3 pixel square presented a few pixels below the center of the screen) and to indicate, via a button press, which adjective seemed a better modifier for the noun. Hand used to indicate “first” or “second” was counterbalanced. A 20-trial practice familiarized subjects with the task. At the start of each trial, four plus signs appeared (500 ms). After a random (800-1300 ms) stimulus onset asynchrony (SOA), the initial adjective was presented centrally (500 ms), followed by 500 ms of blank screen. The noun (subtending 2.5-5.5, mean 4, degrees of visual angle) was then presented (200 ms) with its inner edge two degrees to the left or right of fixation, followed by 1000 ms of blank screen and presentation of the second adjective (500 ms). One second after the offset of the second adjective a question mark appeared, indicating that participants should respond. There were eight blocks of trials, with 49 trials per block. Between blocks, participants took a short break.

The electroencephalogram (EEG) was recorded from 26 silver/silver-chloride electrodes evenly spaced over the scalp. Eye movements were monitored via a bipolar montage of electrodes on the outer canthus of each eye. Blinks were detected by an electrode below the left eye. Signals were amplified with a 0.02-100 Hz bandpass and digitized at 250 Hz. Data were referenced online to the left mastoid and rereferenced offline to the average of the left and right mastoids. Each trial consisted of a 920 ms epoch preceded by a 100 ms prestimulus baseline. Trials contaminated by eye movements, blinks or other recording artifacts were rejected offline. Average trial loss was 4% for the adjectives and 5% for the lateralized nouns. A digital bandpass filter of 0.2 to 20 Hz was employed prior to statistical analyses. Main effects of electrode and interactions with electrode site are reported only when of theoretical significance.



Response accuracy was computed for fillers with one anomalous adjective. In general, responses were quite accurate (mean 84%; range 70-100%), showing that participants were attending to the stimuli and able to apprehend the lateralized nouns. Responses were more accurate in the RVF (88%) than in the LVF (82%) [F(1,31)=8.2, p <.01], consistent with findings of more efficacious word apprehension by the LH (Jordan, 2003).

For critical trials, because both adjectives are related to the noun, both responses are equally correct. Table 2 shows the endorsement rates for the first and second adjectives across conditions and VF. An ANOVA with factors VF (RVF, LVF), Condition Type (Concrete/Concrete, Abstract/Abstract, Concrete/Abstract, Abstract/Concrete), and Adjective Position (first, second) conducted on the endorsement rates revealed no significant effects or interactions. Participants chose concrete and abstract adjectives (LVF/RH: 49%, 51%, RVF/LH: 50%, 50% respectively) and those before and after the critical nouns (LVF/RH: 49%, 51%, RVF/LH: 48%, 52% respectively) in approximately equal proportions.

Table 2
Endorsement rates for the first and second adjectives across conditions and visual fields (C=concrete, A=abstract, U=unrelated, and R=related).


Previous ERP studies have described out of context concreteness effects for nouns (Holcomb et al., 1999; Kounios and Holcomb, 1994; West and Holcomb, 2000) and verbs (Lee and Federmeier, 2008) but not other word classes. Therefore, it was of interest to examine responses to the initial, centrally-presented adjectives (Figure 1). Mean amplitude of the N400 was measured between 300-500 ms over posterior electrodes (marked in Figure 1) and subjected to a repeated measures analysis of variance (ANOVA) with concreteness and electrode as factors. There was a main effect of concreteness [F(1,31)=39.23, p< .001, ηp2= .56], with more negative responses to concrete than abstract adjectives. Effects over frontal sites were evident between 300-500 ms [F(1,31)=34.05, p< .001, ηp2= .52] and 500-900 ms [F(1,31)=18.07, p< .001, ηp2= .37], again in the form of enhanced negativity to more concrete words. Thus, similar to effects seen for nouns and verbs out of context, concrete adjectives elicit larger N400s and a sustained frontal negativity.

Figure 1
Grand average ERPs to concrete (e.g., “green”) and abstract (e.g., “interesting”) adjectives are shown at four electrode sites (left side of the figure). The distribution of the concreteness effect (response difference ...

Lateralized Nouns

Concreteness effects on the lateralized nouns were analyzed for the N400 (posterior channels, 300-500 ms) and frontal negativity (frontal channels, 300-500 ms and 500-900 ms), using ANOVAs with VF (RVF versus LVF), concreteness (concretely versus abstractly modified) and electrode site (11 frontal or 15 posterior) as factors. In addition, analyses were conducted on the frontal P2 (180-240 ms), an ERP component related to high-level visual processing and visuospatial attention, which has proven sensitive to language context effects in a number of VF ERP studies (Federmeier et al., 2005; Wlotko and Federmeier, 2007).

There was a main effect of VF on the P2 [F(1,31)=4.2, p < .05, ηp2= .12], as RVF presentation yielded larger P2 responses. A main effect of concreteness was evident on the frontal negativity (500-900 ms) [F(1,31)=5.9, p < .05, ηp2= .16], with more negative responses to concretely than to abstractly modified nouns. Importantly, there was an interaction of concreteness and VF on the frontal P2, the N400, and the frontal negativity (500-900 ms): F(1,31)=6.59, p < .05, ηp2= .18; F(1,31)=6.06, p < .05, ηp2= .16; and F(1,31)=5.59, p < .05, ηp2= .15, respectively. Follow-up analyses were thus conducted separately in each VF.

Figure 2 shows the ERP responses to concretely and abstractly modified nouns in each VF; for both VFs, the waveforms, timelocked at the noun, immediately follow those to the centrally-presented adjectives shown in Figure 1. Lateralization affected waveform morphology as is typical for VF ERP studies (c.f., Federmeier & Kutas, 1999), such that visual components are larger and peak earlier over contralateral occipital electrode sites and are followed by a long-lasting selection negativity contralateral to presentation VF.

Figure 2
Grand average ERPs to concretely and abstractly modified nouns presented in the RVF/LH (at left) and LVF/RH (at right) are shown at six representative electrode sites (positions indicated by darkened shapes on the head diagram). Arrows highlight the experimental ...

For RVF/LH presentation, concreteness effects were found on the frontal P2 [F(1,31)=6.48; p < .05, ηp2= .17] and the N400 [F(1,31)=7.64; p < .01, ηp2= .2], with more positive responses to concretely than to abstractly modified nouns on both components. The increased positivity over posterior sites to concretely modified nouns continued into the 500-900 ms time window [F(1,31)=8.93; p < .01, ηp2= .22], modulating the amplitude of the late positive complex (LPC), a component that often follows the N400 and has been linked to more explicit aspects of semantic processing (e.g., Swaab et al., 2003). However, no effects were observed on the frontal negativity (F's □ 1). In contrast, with LVF/RH presentation, concreteness effects were significant on the frontal negativity between 500-900 ms [F(1,31)=13.56; p < .001, ηp2= .3], with more negative responses to concretely modified nouns, but not on the frontal P2 or N400 components (F's □ 1). Processing biased to the left and right hemispheres is thus associated with completely distinct types of ERP concreteness effects.

To determine whether the observed effects arise from the concreteness of the unified concept referred to by the adjective-noun pairing, we looked to see whether similar patterns obtained in the unrelated condition, when the concrete or abstract adjective cannot be meaningfully integrated with the noun (e.g., watery glove vs. loyal glove)1. In the RVF/LH, concreteness-related P2 effects were significant even for these meaningless pairs [F(1,31)=7.05; p < .05, ηp2= .19], suggesting that the P2 differences reflect a state induced by a concrete or abstract modifier, irrespective of contextual fit (c.f., Wlotko and Federmeier, 2007). However, N400 and LPC effects in the RVF/LH and the frontal negativity in the LVF/RH were both absent for meaningless adjective-noun pairings (F's □ 1), suggesting that these effects reflect the concreteness of an integrated conceptual representation (Figure 3).


Replicating concreteness effects that have been documented for both nouns (Holcomb et al., 1999; Kounios and Holcomb, 1994; West and Holcomb, 2000) and verbs (Lee and Federmeier, 2008) that are presented out of context, we found that concrete adjectives, presented in sequence-initial positions, elicit more negative ERP responses than do abstract adjectives. Thus, posterior brain responses between 300 and 500 ms (N400) and frontal brain responses between about 300 and 900 ms are consistently different for semantic content picked out by nouns, verbs, and adjectives (e.g. things, actions, and properties) that are more or less associated with sensory features.

Still in dispute, however, are the number and nature of the mechanisms involved in concreteness effects (e.g., whether they are due to differences in the amount of semantic information associated with these words or expressions, to the differential use of mental imagery, or some combination of these) and how these are mediated by lateralized processing resources. Part of the challenge in answering these questions has come from the fact that concreteness effects are normally assessed across different lexical items, which may vary at multiple levels. In this study, therefore, we used adjectives to induce concrete and abstract readings of the same polysemous nouns, thereby holding bottom-up perceptual and lexico-semantic features constant across the concreteness comparison. In order to specifically test the dual coding account of concreteness effects (e.g., Paivio 2007) and, more generally, to examine the dissociability of the processes underlying the electrophysiological signature of these effects, we also lateralized the target nouns to assess hemispheric asymmetries.

Using adjectives to pick out different senses of polysemous nouns was successful in inducing processing differences, which, given that they occurred on exactly same lexical items, can thus be more definitively linked to the concreteness of the evoked concept. Strikingly, these concreteness effects are quite different – indeed, nonoverlapping – in the two hemispheres. With initial presentation to the LH, concreteness effects manifest on the frontal P2 and the N400, two components that have been linked to effects of contextual constraint and target expectancy in language processing (e.g., Wlotko and Federmeier, 2007). In contrast, when processing is biased to the RH, concrete modification leads to a sustained frontal negativity – the part of the ERP concreteness effect that has been associated with mental imagery (West and Holcomb, 2000).

With RVF/LH presentation, concretely modified nouns elicit an enhanced frontal P2 response and a smaller N400. The P2 is part of the normal visual evoked response and has typically been linked to the detection and analysis of higher-level visual features, guided by attention (Luck and Hillyard, 1994; for an alternative perspective see Bles, Alink, & Jansma, 2007). In sentence processing paradigms, the amplitude of the frontal part of the P2 has been found to be modulated by contextual constraint with RVF/LH – but not LVF/RH – processing: targets in strongly constrained sentence contexts (i.e., contexts that elicit strong, consistent predictions for particular endings) elicit more positive P2 responses than those in weakly constrained contexts (Federmeier et al., 2005; Wlotko and Federmeier, 2007). Thus, being able to form strong context-based expectations for upcoming words seems to change how the perceptual processing system allocates attention and analyzes subsequent stimuli. Strikingly, however, these constraint-related P2 enhancements occur not only for the expected lexical item but also for unexpected ones, suggesting that this effect arises at a point in processing prior to the actual identification of the target word (Wlotko and Federmeier, 2007). Similar to the pattern seen in sentences, in the present study the P2 was enhanced for nouns preceded by a concrete modifier irrespective of whether the adjective-noun pair was congruent or anomalous. Thus, for the LH, it seems that concrete adjectives may establish a more constraining context for upcoming nouns than do abstract adjectives.

If, as suggested by the P2 results, concrete adjectives establish more constraining semantic contexts for upcoming nouns, then the large literature on the N400 component (as reviewed in Kutas et al., 2007) would predict reduced N400 amplitudes to nouns following concrete than abstract adjectives, given that increased level of expectancy. Consistent with that prediction, we found that concretely (as opposed to abstractly) modified nouns that are presented initially to the LH facilitate N400 responses, but – distinct from the pattern seen for the P2 – only for congruent pairs. Thus, congruent nouns following a concrete adjective seem to be more expected and thus easier to process semantically. The fact that these expectancy-related effects arise only after presentation to the LH is consistent with theories that posit a specialized role for the LH in using language context information to predict features of upcoming words (the PARLO framework; Federmeier, 2007).

The effect of concrete and abstract modification on the N400 elicited after RVF/LH presentation of the critical noun is different from the pattern seen for concrete and abstract lexical items presented out of context, in which larger N400s have typically been reported for concrete as opposed to abstract words (as was found for the adjectives in the present study). However, the pattern is consistent with prior work suggesting that N400 concreteness effects are sensitive to context (e.g., Holcomb et al., 1999). Across the literature, the pattern of effects suggests that concrete lexical items engender more activity during semantic access when encountered out of context, perhaps because concrete words tend to have more associated feature information (e.g., Kounios & Holcomb 1994), which might be expected to increase baseline N400 amplitudes (see review in Federmeier and Laszlo, in press). However, placing concrete and abstract words into supportive contexts that tend to preactivate many of the target words' features should reduce the magnitude of this difference, as Holcomb et al. (1999) have shown. Furthermore, when concrete and abstract words are themselves used as context for upcoming words, the richer semantic feature information elicited by concrete words (as for the adjectives in the present study) can actually provide a more constraining context for accessing related features of upcoming words. The increased expectancy created by these more constraining contexts would then be expected to lead to N400 amplitude reductions, as seen in the present data for the same target nouns preceded by concrete and abstract adjectives.

The LH, then, seems to treat concrete and abstract words in a qualitatively similar fashion, with differences arising only on the amplitude of components that are obligatorily elicited by all visually-presented content words. A very different pattern, however, is seen when processing of the same stimuli is biased to the RH. In that case, nouns used in a concrete sense elicit a sustained frontal negativity. As discussed, similar sustained frontal negativities have been observed in prior studies of concreteness, as well as studies of mental imagery (Farah et al., 1989)2. In past studies, concreteness effects over frontal sites have often been found to onset earlier (~300 ms) than those in the present study, which were observed between 500 and 900 ms. However, in those studies overlap with the frontal part of the concurrent N400 concreteness effect made the true onset of the sustained frontal negativity difficult to assess. The present results support West and Holcomb's (2000) findings that the frontal effect specifically related to imagery (which they term the “N700”) onsets later.

Importantly, in the present study, this effect arose without instructions or task demands that would bias participants to image, thus suggesting that this process is naturally elicited by the RH when language stimuli are read for meaning, as long as the words can be linked to a coherent conceptual representation. In support of the dual coding hypothesis (Paivio, 2007), the current findings provide a critical link between data suggesting that sensory imagery processes play a role in concreteness effects (Paivio, 1991) and data pointing to RH contributions to the processing of concrete words (Day, 1977; Jones-Gotman and Milner, 1978; Schack et al., 2003; Shallice and Warrington, 1975). Our data are consistent with those of a recent magnetoencephalographic (MEG) study, which found concreteness-related activations between 550 and 700 ms in RH occipital and parietal areas, which the authors linked to imagery (Dhond et al., 2007). The findings of the present study not only support a link between concreteness effects and RH brain areas and processes involved in imagery, but further suggest that the involvement of the RH in language processing is critical for initiating such imagery, as the frontal negativity was not observed when processing was biased toward the LH.

The present data thus provide strong evidence that concreteness effects arise at multiple processing stages mediated by distinct cognitive and neural mechanisms. The LH, well-known to be critical for language processing, clearly derives meaning from both concrete and abstract words and seems to do so in qualitatively similar ways that are nonetheless sensitive to factors that render these word types more or less constrained, expected, or otherwise easy or difficult to process in different contexts. Concrete words and concepts, however, also seem to elicit qualitatively different processing, in the form of mental imagery that creates the rich sensory experience that often accompanies language comprehension. And, strikingly, it is the RH that appears to be critical for mediating this uniquely human ability to mentally recreate kings and horses from words created several centuries in the past.


This research was supported by an NSC GSSA grant to H-W.H and NIA grant AG026308 to K.D.F.




Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1Concreteness values for the adjectives alone were collected in a norming study (N=46) using the same procedures as for the adjective-noun pairs. Mean concreteness ratings were 5.0 for concrete adjectives and 2.6 for abstract adjectives.

2Farah et al. found a slow sustained positivity with a focus over occipital and posterior temporal regions associated with imagery. However, they used Fpz as their reference rather than the more traditional mastoid reference employed in the current study. When the present data are rereferenced to Fpz, the effect observed with LVF/RH presentation shows the distributional pattern of Farah et al.


  • Belmore SM, Yates JM, Bellack DR, Jones SN, Rosenquist SE. Drawing inferences from concrete and abstract sentences. Journal of Verbal Learning and Verbal Behavior. 1982;21:338–351.
  • Binder JR, Westbury CF, McKiernan KA, Possing ET, Medler DA. Distinct brain systems for processing concrete and abstract concepts. J Cogn Neurosci. 2005;17:905–917. [PubMed]
  • Bles M, Alink A, Jansma BM. Neural aspects of cohort-size reduction during visual gating. Brain Research. 2007;1150:143–154. [PubMed]
  • Chiarello C, Liu S, Shears C. Does global context modulate cerebral asymmetries? A review and new evidence on word imageability effects. Brain Lang. 2001;79:360–378. [PubMed]
  • Day J. Right-hemisphere language processing in normal right-handers. J Exp Psychol Hum Percept Perform. 1977;3:518–528. [PubMed]
  • Dhond RP, Witzel T, Dale AM, Halgren E. Spatiotemporal cortical dynamics underlying abstract and concrete word reading. Hum Brain Mapp. 2007;28:355–362. [PMC free article] [PubMed]
  • Ehrlichman H, Barrett J. Right hemispheric specialization for mental imagery: a review of the evidence. Brain Cogn. 1983;2:55–76. [PubMed]
  • Eviatar Z, Menn L, Zaidel E. Concreteness: nouns, verbs, and hemispheres. Cortex. 1990;26:611–624. [PubMed]
  • Farah MJ, Weisberg LL, Monheit MA, Peronnet F. Brain activity underlying mental imagery: Event-related potentials during mental image generation. J Cogn Neurosci. 1989;1:302–316. [PubMed]
  • Federmeier KD. Thinking ahead: the role and roots of prediction in language comprehension. Psychophysiology. 2007;44:491–505. [PMC free article] [PubMed]
  • Federmeier KD, Kutas M. Right words and left words: Electrophysiological evidence for hemispheric differences in meaning processing. Cog Brain Res. 1999;8:373–392. [PubMed]
  • Federmeier KD, Laszlo S. Time for meaning: Electrophysiology provides insights into the dynamics of representation and processing in semantic memory. In: Ross BH, editor. Psychology of Learning and Motivation. Vol. 51. Elsevier; Holland: In press.
  • Federmeier KD, Mai H, Kutas M. Both sides get the point: hemispheric sensitivities to sentential constraint. Mem Cognition. 2005;33:871–886. [PubMed]
  • Fiebach CJ, Friederici AD. Processing concrete words: fMRI evidence against a specific right-hemisphere involvement. Neuropsychologia. 2003;42:62–70. [PubMed]
  • Ganis G, Kutas M, Sereno MI. The search for “common sense”: An electrophysiological study of the comprehension of words and pictures in reading. J Cogn Neurosci. 1996;8:89–106. [PubMed]
  • Gerhand S, Barry C. When does a deep dyslexic make a semantic error? The roles of age-of-acquisition, concreteness, and frequency. Brain Lang. 2000;74:26–47. [PubMed]
  • Glenberg AM, Robertson DA. Symbol grounding and meaning: A comparison of high-dimensional and embodied theories of meaning. J Mem Lang. 2000;43:379–401.
  • Holcomb PJ, Kounios J, Anderson JE, West WC. Dual-coding, context-availability, and concreteness effects in sentence comprehension: an electrophysiological investigation. J Exp Psychol Learn Mem Cogn. 1999;25:721–742. [PubMed]
  • Jones-Gotman M, Milner B. Right temporal-lobe contribution to image-mediated verbal learning. Neuropsychologia. 1978;16:61–71. [PubMed]
  • Jordan TR, Patching GR, Thomas SM. Assessing the Role of Hemispheric Specialisation, Serial-Position Processing, and Retinal Eccentricity in Lateralised Word Recognition. Cognitive Neuropsychology. 2003;20:49–71. [PubMed]
  • Kiehl KA, Liddle PF, Smith AM, Mendrek A, Forster BB, Hare RD. Neural pathways involved in the processing of concrete and abstract words. Hum Brain Mapp. 1999;7:225–233. [PubMed]
  • Kosslyn SM. Seeing and imagining in the cerebral hemispheres: a computational approach. Psychol Rev. 1987;94:148–175. [PubMed]
  • Kounios J, Holcomb PJ. Concreteness effects in semantic processing: ERP evidence supporting dual-coding theory. J Exp Psychol Learn Mem Cogn. 1994;20:804–823. [PubMed]
  • Kutas M, Federmeier KD, Staab J, Kluender R. Language. In: Cacioppo J, Tassinary L, Berntson G, editors. Handbook of Psychophysiology. 3rd edition Cambridge University Press; Cambridge: 2007. pp. 555–580.
  • Lakoff G, Johnson M. Philosophy in the flesh: the embodied mind and its challenge to Western thought. Basic Books; New York: 1999.
  • Lee CL, Federmeier KD. 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. 2008;104:145–158. [PMC free article] [PubMed]
  • Luck SJ, Hillyard SA. Electrophysiological correlates of feature analysis during visual search. Psychophysiology. 1994;31:291–308. [PubMed]
  • Noppeney U, Price CJ. Retrieval of abstract semantics. Neuroimage. 2004;22:164–170. [PubMed]
  • Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia. 1971;9:97–113. [PubMed]
  • Paivio A. Dual coding theory: Retrospect and current status. Can J Psychol. 1991;45:255–287.
  • Paivio A. Mind and its evolution: a dual coding theoretical approach. L. Erlbaum Associates; Mahwah, NJ: 2007.
  • Pexman PM, Hargreaves IS, Edwards JD, Henry LC, Goodyear BG. Neural correlates of concreteness in semantic categorization. J Cogn Neurosci. 2007;19:1407–1419. [PubMed]
  • Pustejovsky J. The Generative Lexicon. MIT Press; Cambridge, MA: 1995.
  • Pylyshyn ZW. Computation and cognition: toward a foundation for cognitive science. MIT Press; Cambridge, MA: 1984.
  • Sabsevitz DS, Medler DA, Seidenberg M, Binder JR. Modulation of the semantic system by word imageability. Neuroimage. 2005;27:188–200. [PubMed]
  • Schack B, Weiss S, Rappelsberger P. Cerebral information transfer during word processing: where and when does it occur and how fast is it? Hum Brain Mapp. 2003;19:18–36. [PubMed]
  • Schwanenflugel PJ. Why are abstract concepts hard to understand? In: Schwanenflugel PJ, editor. The Psychology of word meanings. Lawrence Erlbaum Associates; Hillsale, N.J.: 1991. pp. 223–250.
  • Schwanenflugel PJ, Harnishfeger KK, Stowe RW. Context availability and lexical decisions for abstract and concrete words. J Mem Lang. 1988;27:499–520.
  • Schwanenflugel PJ, Shoben EJ. Differential context effects in the comprehension of abstract and concrete verbal materials. J Exp Psychol Learn Mem Cogn. 1983;9:82–102.
  • Shallice T, Warrington EK. Word recognition in a phonemic dyslexic patient. Q J Exp Psychol. 1975;27:187–199. [PubMed]
  • Swaab T, Brown C, Hagoort P. Understanding words in sentence contexts: the time course of ambiguity resolution. Brain Lang. 2003;86:326–343. [PubMed]
  • Swaab T, Baynes K, Knight RT. Separable effects of priming and imageability on word processing: an ERP study. Brain Res. 2002;15:99–103. [PubMed]
  • Toglia MP, Batting WF. Handbook of Semantic Word Norms. Lawrence Erlbaum; Hillsdale, NJ: 1978.
  • West WC, Holcomb PJ. Imaginal, semantic, and surface-level processing of concrete and abstract words: an electrophysiological investigation. J Cogn Neurosci. 2000;12:1024–1037. [PubMed]
  • Wlotko EW, Federmeier KD. Finding the right word: hemispheric asymmetries in the use of sentence context information. Neuropsychologia. 2007;45:3001–3014. [PMC free article] [PubMed]