Anticipating words and their gender: an event-related brain potential study of semantic integration, gender expectancy, and gender agreement in Spanish sentence reading

Abstract

Recent studies indicate that the human brain attends to and uses grammatical gender cues during sentence comprehension. Here, we examine the nature and time course of the effect of gender on word-by-word sentence reading. Event-related brain potentials were recorded to an article and noun, while native Spanish speakers read medium- to high-constraint Spanish sentences for comprehension. The noun either fit the sentence meaning or not, and matched the preceding article in gender or not; in addition, the preceding article was either expected or unexpected based on prior sentence context. Semantically anomalous nouns elicited an N400. Gender-disagreeing nouns elicited a posterior late positivity (P600), replicating previous findings for words. Gender agreement and semantic congruity interacted in both the N400 window--with a larger negativity frontally for double violations--and the P600 window--with a larger positivity for semantic anomalies, relative to the prestimulus baseline. Finally, unexpected articles elicited an enhanced positivity (500-700 msec post onset) relative to expected articles. Overall, our data indicate that readers anticipate and attend to the gender of both articles and nouns, and use gender in real time to maintain agreement and to build sentence meaning.

Figure 1

Representative grand average ERPs from eight electrode sites (indicated by the X's on the schematic head) from front to the back of the head to the noun targets. On this and all subsequent figures, negative is plotted upward and time is represented on the x-axis in milliseconds. The left half of the figure shows the effect of semantic congruity; overlapped is the ERP to nouns that made sense in the sentence context (solid line) versus the ERP to those that did not (dashed line). Note greater negativity to the incongruous nouns relative to the congruous ones from 200 msec throughout the recording epoch. The right half of the figure shows the effect of gender agreement between the article and noun; overlapped is the response to nouns when it agrees in gender with that of the prior article (solid) versus when it does not (dashed line). Note that gender mismatches are associated with a greater positivity from 400 msec throughout the recording epoch.

Figure 2

Grand average ERPs to target nouns from all 26 electrode sites (looking down on top of the head) with the four experimental conditions overlapped. Note the bilateral distribution of the increased frontal negativity to the double violation condition (semantic and gender violations) compared with the semantic violation alone, as well as the LPC distribution over posterior sites for all violations relative to the control (semantically congruous and gender matching). The two highlighted electrodes (left medial frontal and left medial occipital) appear enlarged in Figure 3.

Figure 3

Grand average ERPs to target nouns for the 4 experimental conditions overlapped, shown at an anterior and a posterior site relative to two different baselines. On the left, relative to a 100-msec prestimulus baseline, the largest negativity in the N400 region is elicited by the double violation at the anterior site, while the beginning of the LPC is visible at the posterior site. In the LPC/P600 region, the interaction differs at anterior and posterior sites, but relative to the prestimulus baseline, the positivity is largest for gender-mismatching semantically congruous nouns. On the right, the same data are baselined on the peak of the N400 (i.e., peak-to-peak comparison), thereby taking into account differences in N400 amplitude across conditions. In contrast to the prestimulus baseline comparison, measured peak-to-peak, it is the double violation that elicits the largest positivity in the LPC/P600 region; there was no significant difference between the semantic-violation and gender-violation conditions nor any interaction between the two.

Figure 4

Difference ERPs at 8 electrodes indicated from front to back on the schematic head illustrate the interaction between semantic congruity and gender agreement relative to a prestimulus baseline. The left half of the figure shows that the effect of semantic congruity violations (incongruous minus congruous ERPs) is greater when the noun also mismatches in gender with the preceding article (dashed line, which corresponds to the difference between the dotted lines in the left half of Figure 3) than when it matches (solid line, which corresponds to the difference between the solid lines in Figure 3). In the N400 time window, the congruity effect (incongruous minus congruous) is characterized by relatively greater negativity for gender mismatches relative to gender matches across the scalp; this widespread negativity of the congruity effect for gender mismatches versus matches arises from the contribution of the greater frontal negativity to incongruous gender mismatches and the greater positivity to congruous gender mismatches. The right half of this figure shows that the effect of gender agreement (mismatch minus match) is greater when the noun fits (solid line, corresponding to the difference between dark-solid and light-dotted lines in left half of Figure 3) than when it does not fit with the meaning of the sentence context (dashed line, corresponding to the difference between light-solid and dark-dotted lines in Figure 3). None of the small differences in these difference ERP waveforms is reliable.

Figure 5

Difference ERPs from four midline electrodes (MiPf, MiCe, MiPa, and MiOc) from front to back of the head comparing the double-violation effect (solid line in all three columns: gender-mismatching-semantically incongruous minus matching-congruous ERPs) to the pure effects of semantic congruity (left column: gender matching, semantic incongruous minus congruous) and gender agreement (middle column: semantically congruous, gender-mismatching minus matching). The right column shows that the sum of the effects of pure gender agreement and pure semantic congruity violations does not equal the brain's response to the double violations (two violations presented together).

Figure 6

Grand average ERPs from all 26 electrode sites (looking down on the top of the head) for the effect of gender expectancy at the article preceding the target noun. The effect is characterized by a widely distributed enhanced positivity for the contextually unexpected (dashed line) versus expected (solid line) articles between 500 and 700 msec (here shown only through 650 msec since the visual EP to the following word is too large to graph at this scale; see Figure 7 for the full two word waveforms). None of the smaller differences is reliable.

Figure 7

Comparison across four midline electrodes from front to back of the head of the main effect of gender agreement collapsed across semantic congruity for a three-word time window (one word every 500 msec), including the target article and noun in the current study (far left) and target article and picture (middle) embedded in written sentences (presented one word at a time in Wicha et al.'s, 2003, study). On the far right, the difference ERPs (gender mismatch minus match) across experiments are compared. The effects across the two experiments have a similar time course, but differ in polarity at both the article (onset at 0 msec) and target noun/picture (onset at 500 msec).