Abstract

Neuropsychological research investigating mental grammar and lexicon has largely been based on the processing of regular and irregular inflection. Past tense inflection of regular verbs is assumed to be generated by a syntactic rule (e.g., show-ed), whereas irregular verbs consist of rather unsystematic alternations (e.g., caught) represented as lexical entries. Recent morphological accounts, however, hold that irregular inflection is not entirely rule-free but relies on morphological principles. These subregularities are computed by the syntactic system. We tested this latter hypothesis by examining alternations of irregular German verbs as well as pseudowords using ERPs. Participants read series of irregular verb inflection including present tense, past participle, and past tense forms embedded in minimal syntactic contexts. The critical past tense form was correct (e.g., er sang [he sang]) or incorrect by being either partially consistent (e.g., *er sung [*he sung]) or inconsistent (e.g., *er sing [*he sing]) with the proposed morphological principles. Correspondingly, in a second experimental block, pseudowords (e.g., tang/*tung/*ting) were presented. ERPs for real words revealed a biphasic ERP pattern consisting of a negativity and P600 for both incorrect forms in comparison to the correct equivalents. Most interestingly, the P600 amplitude for the incorrect forms was gradually modulated by the type of anomaly with medium amplitude for consistent past tense forms and largest amplitude for inconsistent past tense forms. ERPs for pseudoword past tense forms showed a similar gradual modulation of N400. The findings support the assumption that irregular verbs are processed by rule-based mechanisms because of subregularities of their past tense inflection.

INTRODUCTION

The processing of inflectional morphology received much attention in neuropsychological research for characterizing the human language system. A seminal view postulates a modular system with specialized modules for computation and retrieval of linguistic information (Marslen-Wilson & Tyler, 2007; Ullman, 2001, 2004). The distinction into different neurocognitive systems is largely based on the assumption that a majority of verbs (i.e., regular verbs) relies on decompositional processing by application of grammatical rules, whereas a minor class of verbs (i.e., irregular verbs1) involves nondecompositional processing by retrieval of their whole forms from the mental lexicon. This view, however, neglects potential regular characteristics underlying the latter class of verbs.

Past tense inflection of most languages is largely characterized by regularity by means of grammatical rules, such as affixation of -ed to the verb stem for English past tense. Despite that, some verbs are resistant to these basic grammatical rules and exhibit irregular inflection by undergoing various alternations, such as alternations of the stem vowel (e.g., sing > sang, know > knew). Thus, irregular inflection has often been described as less predictable and apparently unsystematic. A more fine-grained view, however, suggests that irregular past tense inflection can be covered by abstract inflection patterns, such as ABC, ABB, or AAB,2,3 which describe the different types of stem alternations. Accordingly, the ABC pattern refers to irregular verbs having alternated stem forms for past participle and past tense each (for examples, see Table 1). The other two patterns denote alternated stem forms for both past participle and past tense forms (i.e., ABB), as well as for the past tense form only (i.e., AAB). In various Germanic languages (e.g., English, German, or Dutch), similar inflection patterns co-occur, indicating a cross-linguistic phenomenon. Notably, the pattern of ABA representing an identical stem form for present and past tense, but an alternated one for past participle only, is not observed. In morphological analysis, Wiese (2008) describes these systematic patterns by means of morphological principles.

Table 1. 

Samples of Irregular English (First Column Each), Dutch (Second Column Each), and German (Third Column Each) Verbs Ascribable to the Different Inflection Patterns

Inflection PatternPresentPast ParticiplePast Tense
ABC swim spreken helfen swum gesproken geholfen swam sprak half 
AAB become geven lesen become gegeven gelesen became gaf las 
ABB break binden heben broken gebonden gehoben broke bond hob 
Inflection PatternPresentPast ParticiplePast Tense
ABC swim spreken helfen swum gesproken geholfen swam sprak half 
AAB become geven lesen become gegeven gelesen became gaf las 
ABB break binden heben broken gebonden gehoben broke bond hob 

Underspecification-based Approaches

By morphological analysis of the form–function relation between irregular stem forms, systematic aspects of past tense alternations have been discovered (e.g., Mailhammer, 2008; Wunderlich & Fabri, 1995; Halle & Marantz, 1993; Fabricius-Hansen, 1977). From an underspecification-based perspective,4 stem alternations of irregular verbs serve as morphological markers for syntactic categories, comparable to -ed marking for regular past tense, and rely on morphological rules (Wiese, 2008). According to Wiese's (2008) approach, the morphosyntactic properties of irregular stem forms in present tense, past participle, and past tense are described by an attributed set of abstract feature values (e.g., past tense, finiteness). By carrying distinct feature sets, the different stem forms vary in their specificity. Present tense stem forms are maximally unspecific (i.e., underspecified) and are specified by an empty feature set [ ]. Past participle stem forms, however, are more specific than present tense stem forms and are marked by the feature [+past]. Past tense stem forms are presumed to be most specific and to comprise the features [+past] and [+finite]. The occurrence of particular stem forms is assumed to depend on the principle of specificity,5 stating that whenever more specific stem forms exist, they take priority over less specific forms (cf. Wunderlich, 1996; Wunderlich & Fabri, 1995). This principle functionally accounts the observable inflection patterns: In case of the ABC pattern, for all tense forms gradually specific stem forms exist, allowing a rule-based insertion of past participle and past tense forms. In the ABB pattern, a more specific past tense stem form (i.e., form C carrying the features [+past, +finite]) is absent, so that the next specific form (i.e., form B carrying the feature [+past]) is applied. Following the principle of specificity, an inflection pattern of ABA, however, in which an unspecific stem form A would take precedence over a more specific stem form B in past finite contexts, cannot occur at all. Because of their different morphosyntactic properties and rule-based insertion, the irregular stem forms are therefore presumed to consist of subregularities.

Neuropsychological Approaches on Regular and Irregular Verb Processing

The different notions of regular and irregular verb inflection are reflected in neuropsychological modeling. According to dual-mechanism models, regular past tense forms (e.g., walk + ed) are decomposed into verb stem and affix by an internal rule-based mental grammar (Clahsen, 1999; Pinker, 1999; Marslen-Wilson & Tyler, 1998; Ullman et al., 1997; Schreuder & Baayen, 1995). Irregular past tense forms, however, are stored in the mental lexicon and are retrieved as whole forms during lexical access. These different neurocognitive operations (decomposition vs. unparsed retrieval) have been assumed to underlie distinct neural substrates. A frontal procedural memory system including Broca's area (i.e., Brodmann's area 44 and 45) and the BG is assumed to be responsible for morphophonological analysis processes, whereas an associative mental lexicon subserves the retrieval of lexicalized past tense forms (Marslen-Wilson & Tyler, 2007; Ullman, 2001; Ullman et al., 1997).

In contrast, connectionist approaches propose similar neuronal networks to underlie the processing of both regular and irregular past tense forms, irrespective of whether or not inflected verb forms can be morphophonologically decomposed into stem and affix (Joanisse & Seidenberg, 1999, 2005; Rumelhart & McClelland, 1986). The processing and representation of both types of verbs are assumed to depend on a unitary language system, in which inflected tense forms are represented as overlapping whole forms, which have certain phonological and semantic features in common.

According to underspecification-based accounts, irregular past tense forms are morphophonologically parsed equivalent to regular ones, instead of being accessed as whole forms, by assuming that the different patterns of irregular past tense inflection occur systematically because of morphological rules (Wiese, 2008).6 This account proposes subregularities to underlie irregular inflection, thereby involving morphophonological analysis and computation by the syntactic component of the language system.

ERPs

Evoked brain potentials allow the investigation of neurocognitive processes underlying human language processing by providing time-sensitive measures of the brain activity during stimulus processing. ERP studies investigating morphological and syntactic information processing have shown distinct ERP components, such as LAN and P600. The LAN component emerges around 300 msec after stimulus presentation and shows a left-lateralized or bilateral anterior scalp distribution. LAN effects were seen for various types of morphosyntactic anomalies, including disagreement of subject–verb relations and tense inflection (e.g., O'Rourke & Van Petten, 2011; Newman, Ullman, Pancheva, Waligura, & Neville, 2007; Hagoort, Wassenaar, & Brown, 2003; Gunter, Friederici, & Schriefers, 2000; Coulson, King, & Kutas, 1998; Münte, Heinze, & Mangun, 1993; Rösler, Putz, Friederici, & Hahne, 1993; Neville, Nicol, Barss, Forster, & Garrettdew, 1991; Kutas & Hillyard, 1983). LANs have also been elicited by morphological anomalies, comprising violations of morphological structure (Lück, Hahne, & Clahsen, 2006; Rodriguez-Fornells, Clahsen, Lleo, Zaake, & Münte, 2001; Gross, Say, Kleingers, Clahsen, & Münte, 1998; Penke et al., 1997; Weyerts, Penke, Dohrn, Clahsen, & Munte, 1997; Friederici, Pfeifer, & Hahne, 1993). For overregularization of regular inflection to irregular nouns (e.g., *Pantoffel-s vs. Pantoffel-n), Weyerts and colleagues (1997) reported an enhanced LAN, whereas no such ERP effect was seen for irregularization of regular nouns (e.g., *Trick-en vs. Trick-s). Lück and colleagues (2006) replicated these findings for the auditory modality, indicating that violations of morphological constraints reliably evoked LAN effects. Although LANs are typically anteriorly distributed, more widespread distribution of LAN has also been reported (Hasting & Kotz, 2008; Kaan & Swaab, 2002; Friederici & Frisch, 2000; Gross et al., 1998). In response to a violation of subject–verb agreement in German (e.g., *er kegel-st vs. er kegel-t), a widespread negativity extending to anterior and posterior sites was elicited (Hasting & Kotz, 2008). The LAN has been interpreted as an index of syntactic–relational processing (Friederici, 2002) as well as a reflection of parsing processes associated with morphological structure building (e.g., Ullman, 2004; Weyerts et al., 1997).

For violations of morphosyntactic constraints, a late positivity (P600), subsequent to LAN, has often been observed. The P600 emerges around 500 msec poststimulus onset and has a centroparietal scalp distribution. This late positivity is reliably evoked by a variety of syntactic anomalies, including violations of phrase structure (Friederici & Meyer, 2004; Hagoort, Brown, & Groothusen, 1993), verb argument structure (Friederici & Frisch, 2000; Osterhout & Hagoort, 1999), or morphosyntactic constraints (e.g., Newman et al., 2007; Hagoort, 2003; Münte, Matzke, & Johannes, 1997). Late positivities in response to syntactic anomalies were, thus, also referred to as syntactic positive shift (Hagoort et al., 1993; Osterhout & Holcomb, 1992). With regards to the functional significance of the P600, this component has been associated with repair and reanalysis processes of the anomalous sentence structure (Friederici, 2002). The function of P600 is still controversially debated, since this brain potential has shown to be sensitive to semantic–2011; Regel, Gunter, & Coulson, 2010) as well as nonlinguistic information processing (Coulson et al., 1998). Recent findings imply distinct neurocognitive processes (i.e., syntactic vs. pragmatic reanalysis) to be reflected in the P600 (Regel, Meyer, & Gunter, 2014).

Besides, several studies have reported an N400 component, a centroparietal negativity with a peak latency of around 400 msec after stimulus onset, in response to morphological anomalies (Leinonen, Brattico, Jarvenpaa, & Krause, 2008; Hahne, Müller, & Clahsen, 2006; Morris & Holcomb, 2005). Because the N400 is mainly known to reflect lexical–semantic processes, observations of an N400 for irregularizations of regular noun plurals (e.g., *Waggon-en vs. Waggon-s) have been claimed to reflect enhanced lexical processing with the irregularized plural forms being treated as nondecomposable units, which are processed like pseudowords (Lück et al., 2006; Weyerts et al., 1997). N400 effects have been reliably found for pseudowords exhibiting some similarities to existing words (e.g., PANK vs. BANK), suggesting that the N400 is associated with lexical access (Laszlo & Federmeier, 2009; Lehtonen et al., 2007). Moreover, the amplitude of N400 was also modulated by the degree of prediction for an upcoming sentence completion that either matches or mismatches expected syntactic and orthographic characteristics (Laszlo & Federmeier, 2009; DeLong, Urbach, & Kutas, 2005).

In previous ERP studies, the processing of regular and irregular past tense inflection has been examined for various languages, including English (Newman et al., 2007; Morris & Holcomb, 2005; Allen, Badecker, & Osterhout, 2003; Münte, Say, Clahsen, Schiltz, & Kutas, 1999), German (Penke et al., 1997), Spanish (de Vega, Urrutia, & Dominguez, 2010; Linares, Rodriguez-Fornells, & Clahsen, 2006; Rodriguez-Fornells, Münte, & Clahsen, 2002), Italian (Gross et al., 1998), or Catalan (Rodriguez-Fornells et al., 2001). In a study by Penke and colleagues (1997), correctly and incorrectly inflected German past participles were compared. Incorrect irregular participles (i.e., because of addition of the regular affix to the verb stem [e.g., *gelad-et vs. gelad-en]) elicited a LAN with a left frontotemporal scalp distribution relative to correct inflections. Furthermore, in a study by Morris and Holcomb (2005), a LAN–P600 pattern emerged for overregularizations of irregular verbs by affixation of -ed to the verb stem (e.g., *bringed vs. brought) as well as for irregularizations of regular verbs by application of stem alternations (e.g., *sept vs. seeped). The LAN observed for irregularizations was assumed to reflect difficulty in retrieving syntactic information from words having no lexical entry. Gross and colleagues (1998) also reported a negativity with a more widespread distribution for overregularized Italian past participles (e.g., *prendato vs. preso). For overregularized irregular Catalan past tense forms consisting of violations of stem formation rules (e.g., *dorm-a-t vs. dorm-i-t), Rodriguez-Fornells and colleagues (2001) also observed a left lateralized negativity suggesting a sensitivity of LAN to morphological structure building processes.

The Present Study

The present ERP study examines systematic paradigms of irregular past tense inflection and aims to provide further evidence for neuropsychological modeling on language processing. Two different types of incorrect German irregular past tense forms (i.e., partially consistent vs. inconsistent forms) are compared to grammatically correct forms to scrutinize whether their processing differs systematically according to the proposed morphosyntactic properties (see Table 2). Participants were presented with series of two-word and three-word items inflected for different tenses (i.e., future tense, past participle, and past tense). Critical past tense items were correct or incorrect by application of either the past participle stem (i.e., the consistent condition) or the present tense stem (i.e., the inconsistent condition)7 (for an example, see Table 2). If irregular past tense forms are morphophonologically parsed, a LAN component would be expected for both incorrect forms relative to the correct equivalent. In addition, a subsequent P600 component is predicted for incorrect past tense forms. Observing a LAN–P600 pattern for violation of irregular past tense inflection would indicate that irregular verbs rely on a rule-based processing, comparable to regular verbs. Moreover, if irregular past tense inflection is indeed based on morphological principles (cf. Wiese, 2008), a gradual modulation of P600 is predicted. For consistent past tense forms partially agreeing with relevant morphosyntactic properties of the correct past tense form (i.e., sharing the feature [+past]), smaller P600 amplitude is expected than for inconsistent past tense forms lacking context relevant morphosyntactic properties (i.e., the features [+past] and [+finite]). This finding would suggest that the different morphosyntactic properties underlying irregular past tense forms are taken into account for reanalysis of the anomalous item structure, thereby confirming the existence of morphological principles within irregular verb inflection. If, however, irregular past tense forms are stored in the mental lexicon and involve non-rule-based processing, as proposed by both the dual mechanism models and connectionist accounts, then ERPs should only reveal an N400 effect for incorrect compared to correct past tense forms. This finding would imply that potential subregularities of irregular inflection do not exist.

Table 2. 

Systematic Manipulation of Irregular Past Tense Forms according to the Morphological Analysis of Wiese (2008), Exemplarily Shown for the Verb Singen (Sing)

ConditionPresentPast ParticiplePast TenseMorphosyntactic PropertiesInflection Pattern
correct er wirdsingen er hat gesungen er sang [+past, +finite] ABC 
incor: consistent er wirdsingen er hat gesungen *er sung [+past] ABB 
incor: inconsistent er wirdsingen er hat gesungen *er sing [ ] *ABA 
ConditionPresentPast ParticiplePast TenseMorphosyntactic PropertiesInflection Pattern
correct er wirdsingen er hat gesungen er sang [+past, +finite] ABC 
incor: consistent er wirdsingen er hat gesungen *er sung [+past] ABB 
incor: inconsistent er wirdsingen er hat gesungen *er sing [ ] *ABA 

Present and participle stem forms are underlined.

incor = incorrect.

A related question of this study was whether the proposed morphological principles underlying irregular inflection depend on the lexical representation of the verbs, or whether they denote rather general principles. To examine this issue, irregular verbs were replaced by pseudowords in a separate experimental block, subsequent to the real word presentation. Pronounceable pseudowords were created in analogy to morphophonological structures of existing irregular verbs and embedded into two- and three-word items (e.g., er wird tingen, er hat getungen, er tang/*tung/*ting). Thus, the supposed grammaticality of the pseudowords corresponded to that of the real words (i.e., correct, consistent, and inconsistent). If the assumed subregularities of irregular past tense inflection reflect principled morphological rules, incorrect past tense pseudowords should differ from equivalent correct ones. In case morphological rules are applied to the pseudoword items and allow the participants to predict morphophonological characteristics of an upcoming past tense form, ERPs should reveal a prediction-related N400 component for the incorrect pseudoword items. Moreover, the N400 amplitude is expected to gradually modulate with smaller amplitude for consistent pseudoword items and larger amplitude for inconsistent pseudoword items. If ERPs, however, show no differences between the correct and incorrect pseudoword items, such a finding would suggest that potential subregularities of irregular verb inflection depend on the lexical representation of verbs and would not reflect a general principle.

METHODS

Participants

In the experiment, 42 native German-speaking students (21 women, mean age = 24.5 years, SD = 2.75) participated. All participants were right-handed, had a normal or corrected-to-normal vision, and were paid for their participation.

Stimulus Material

The stimulus material consisted of 36 existing German irregular verbs as well as 36 corresponding pronounceable pseudowords. The irregular verbs belonged to the ABC inflection pattern and comprised stem alternations for the past participle and past tense forms. Half of the verbs were chosen from the /e-o-a/ inflection series, and half stemmed from the /i-u-a/ series. The full list of materials can be found in Appendix A. The past tense forms were mainly monomorphemic (mean number of syllables = 1.17, SD = 0.38; mean word length = 5.14, SD = 1.13) and had an average frequency class of 12.9 (SD = 3.52) according to the Leipzig vocabulary project (www.wortschatz.uni-leipzig.de). All verbs were inflected for the three different tenses and embedded in minimal syntactic contexts, thereby creating three-word and two-word items (e.g., er sprach [he spoke], es sank [it sank]). The past tense forms were systematically manipulated, yielding either grammatically correct items (i.e., correct condition) or incorrect items (i.e., consistent and inconsistent condition). For the consistent condition, the past tense form was replaced by the past participle stem form (e.g., *er sproch [he spoke], *es sunk [it sunk]). For the inconsistent condition, the present tense stem form was applied instead of the past tense form (e.g., *er sprech [he speak], *es sink [it sink]). Hence, the critical information was contained in the past tense item. Note that the items did not differ in phonological complexity but varied only in phonological feature specification (i.e., [+/− high] vs. [+/− high], [+ front]). An experimental item consisted of three inflected items presented in a series, each beginning with future tense, followed by past perfect and past tense (e.g., er wird sprechen [he will speak], er hat gesprochen [he has spoken], er sprach [he spoke]). In total, 108 inflected verb series with 36 items per condition were created. To match the number of grammatically correct items with that of incorrect items, 12 irregular grammatically correct filler verbs that stemmed from the inflection series of /ei-i-i/ were included (e.g., er wird streiten, er hat gestritten, er stritt [he quarreled]). The filler items were similar to the real word items and presented in inflected series.

For the pseudoword block, 36 equivalent pronounceable pseudowords and 12 equivalent pseudoword fillers were created (e.g., binken, preiten) and embedded in existing minimal syntactic contexts (e.g., er wird pfechen, er hat gepfochen, er pfach). The full list of pseudoword materials can be found in Appendix B. The pseudoword items were presented likewise the real word items in inflected verb series comprising the above-mentioned tense forms. Note that the pseudoword verbs were as similar as possible to the real word verbs in terms of their syllable structure and word length.

For each experimental block (i.e., the real word block, and the pseudoword block), the factor inflection (correct/consistent/inconsistent) was applied.

Procedure

Participants were tested individually in a 25-min session during which they were seated in a soundproof, dimly lit cabin. Participants were asked to read attentively all verb series and to reply as accurately as possible to the experimental task with a yes or no response. For the real word block, participants had to judge whether the past tense item was grammatically correct or not. For the pseudoword block, participants were asked to judge whether the pseudoword past tense item was appropriate within a particular series of verb forms, or not. In each block, half of the trials were correct and half were incorrect. All responses were given via button press. Before each experimental block, participants received an instruction and a short training phase of 10 trials.

For experimental presentation, real word trials and pseudoword trials were presented in separate experimental blocks, whereby the real word block always preceded the pseudoword block. For each block, the stimulus material was pseudorandomized and divided into three item versions with 48 items each (i.e., 36 experimental items and 12 filler items). In this way, each experimental item was only presented once within each version of each block. Experimental conditions were equally divided across all versions (i.e., 12 items of each condition). A trial sequence started off with the presentation of a fixation cross for 200 msec in the middle of the monitor. After an ISI of 300 msec, item series were presented visually with 1500 msec for each item and a 500-msec pause in between (a rate that was comfortable for participants). An item series began with the future tense item, followed by the past perfect and past tense item. All words of an item appeared at once on the monitor. Participants were instructed to avoid eye movements during the presentation of the items. After offset of the stimulus presentation and an additional interval of 1500 msec, participants had to perform the experimental task (i.e., grammaticality judgment for the real word block, and appropriateness judgment for the pseudoword block). After the response was given (RT of maximal 3000 msec) and an intertrial interval of 1000 msec, the next trial started. Yes and no answers were completely balanced across all experimental conditions in avoidance of a decision-related expectancy.

Data Recording and Analysis

Behavioral data consisted of the grammaticality judgments as well as the appropriateness judgments obtained in the two experimental blocks, which were analyzed separately in a repeated-measures ANOVA with the factor Inflection (correct/consistent/inconsistent). To analyze effects of Inflection, post hoc t test for pairwise comparison was employed. Because of multiple comparisons, all effects observed in post hoc testing are corrected by the Bonferroni–Holm procedure.

ERP data were extracted from the EEG recorded from 52 Ag–AgCl electrodes.8 During the signal recording, the EEG was referred to the right mastoid and afterwards re-referenced to the average of the left and right mastoids. The recording included bipolar horizontal and vertical EOG. Resistance of all electrodes was kept under 5 KΩ. EEG and EOG signals were recorded continuously with a band pass between DC and 70 Hz and a sampling rate of 500 Hz. Average ERPs were computed for the critical item (i.e., the past tense item) for each electrode position for each of the six experimental conditions. Averages were aligned to a 200-msec prestimulus baseline and calculated for a period of 1000 msec after stimulus onset. For the real word block, statistical analysis included only correctly answered trials. For the pseudoword block, each trial entered the analysis. Approximately 5% of all trials were excluded from the averages because of ocular artifacts (automatic EOG rejection ±40 μV).

For statistical analysis of the ERP data, three latency windows were employed: 250–400 msec (LAN) and 400–800 msec (P600) for real word items and 300–500 msec (N400) for pseudoword items. Multivariate analyses of variance were performed separately for the two experimental blocks on all dependent variables. To avoid problems concerning sphericity, the multivariate approach to repeated measurements was applied (Dien, Spencer, & Donchin, 2004; Vasey & Thayer, 1987). For distributional ERP analyses, two topographical factors Anterior/posterior and Hemisphere (left [LH]/right [RH]) were defined and completely crossed, yielding four different ROIs containing eight electrodes each: left anterior (F5, F3, Ft7, Fc5, Fc3, T7, C5, C3), left posterior (Tp7, Cp5, Cp3, P7, P5, P3, Po7, Po3), right anterior (F4, F6, Fc4, Fc6, Ft8, C4, C6, T8), and right posterior (Cp4, Cp6, Tp8, P4, P6, P8, Po4, Po8). Within-subject factors were Anterior/posterior (2), Hemisphere (2), and Inflection (correct/consistent/inconsistent). Whenever interactions between two or more factors were found, further analyses were carried out. Midline electrode positions (Fz, Fcz, Cz, Cpz, Pz, Poz) were analyzed separately. Post hoc tests using t test for pairwise comparison were applied to analyze main effects of Inflection. All effects thereby observed are corrected by the Bonferroni–Holm procedure.

RESULTS

Behavioral Data

For the real word block, participants' performance on the grammaticality judgment was excellent and revealed a mean accuracy rate of 98.5% (SD = 3.64). Statistical analysis showed a main effect of Inflection, F(2, 82) = 16.22, p < .001. Further analyses revealed significant differences between the correct and consistent condition, F(1, 41) = 21.50, p < .0001, as well as between the consistent and inconsistent condition, F(1, 41) = 17.66, p < .0001. This finding suggests that for the consistent condition more errors were made (2.8% errors) than for the correct (0.5% errors) as well as the inconsistent condition (1.0% errors).

In the pseudoword block, an average accuracy rate of 94.6% (SD = 9.95) was found for the appropriateness judgment of pseudoword past tense forms indicating a very good performance of the participants. Statistical analysis showed a main effect of Inflection, F(2, 82) = 43.02, p < .0001. Multiple comparisons between single conditions revealed significant differences between the correct and consistent condition, F(1, 41) = 51.04, p < .0001, as well as between the consistent and inconsistent condition, F(1, 41) = 58.12, p < .0001. The findings suggest that the consistent pseudoword past tense forms could less accurately be judged (error rate of 10.5%) than the correct (error rate of 2.9%) and the inconsistent past tense forms (error rate of 2.6%).

Electrophysiological Data

ERPs for the Real Word Block

Grand average ERPs at the critical item suggested the elicitation of a negativity around 300 msec poststimulus onset and a later P600 for both incorrect past tense items in comparison to the grammatically correct equivalents. Most interestingly, the amplitude of P600 was gradually modulated by specificity with medium amplitude for consistent items and largest amplitude for inconsistent items (Figure 1).

Figure 1. 

Grand average ERPs of the real word past tense items that were grammatically correct (blue line), consistent (green line), and inconsistent (red line). The zoom of the PZ electrode below illustrates the modulation of the P600 component.

Figure 1. 

Grand average ERPs of the real word past tense items that were grammatically correct (blue line), consistent (green line), and inconsistent (red line). The zoom of the PZ electrode below illustrates the modulation of the P600 component.

The 250–400 msec latency window

In the LAN time window, a main effect of Inflection, F(2, 40) = 5.30, p < .01, was revealed. An interaction of inflection with Anterior/posterior, F(2, 40) = 8.13, p < .001, was significant, and a further interaction of Inflection with Hemisphere, F(2, 40) = 2.71, p < .1, was marginally significant. This latter interaction was resolved by hemisphere and showed significant effects of Inflection for both the LH and RH, F(2, 40) = 4.07–5.64, p < .05. The observed results of post hoc testing are shown in Table 3, in the upper part. The findings imply the presence of a negativity for both incorrect past tense items in relation to grammatically correct items, which was distributed over left and right electrode sites.

Table 3. 

Effects of the Post hoc Tests of the 250–400 msec and 400–800 msec Latency Window Seen for the LH and RH, as well as for the Midline Electrode Sites

EffectLHRHMidline Sites
dftdftdft
Latency Window of 250–400 msec 
correct vs. con 1,41 10.42*** 1,41 7.39*** 1,41 12.55*** 
correct vs. incon 1,41 6.55** 1,41 5.95** 1,41 7.41*** 
con vs. incon 1,41 ns 1,41 ns 1,41 ns 
Latency Window of 400–800 msec 
correct vs. con 1,41 ns 1,41 4.96** 1,41 4.84* 
correct vs. incon 1,41 13.14*** 1,41 12.95*** 1,41 19.43*** 
con vs. incon 1,41 8.17*** 1,41 ns 1,41 5.47** 
EffectLHRHMidline Sites
dftdftdft
Latency Window of 250–400 msec 
correct vs. con 1,41 10.42*** 1,41 7.39*** 1,41 12.55*** 
correct vs. incon 1,41 6.55** 1,41 5.95** 1,41 7.41*** 
con vs. incon 1,41 ns 1,41 ns 1,41 ns 
Latency Window of 400–800 msec 
correct vs. con 1,41 ns 1,41 4.96** 1,41 4.84* 
correct vs. incon 1,41 13.14*** 1,41 12.95*** 1,41 19.43*** 
con vs. incon 1,41 8.17*** 1,41 ns 1,41 5.47** 

con = consistent; incon = inconsistent; ns = not significant.

*p < .05.

**p < .025.

***p < .016.

For midline electrode sites, an effect of Inflection was seen, F(2, 40) = 6.48, p < .01. The results confirm the emergence of a negativity for incorrect relative to correct items that was also present over midline electrode sites.

The 400–800 msec latency window

In the P600 time window, a main effect of Inflection, F(2, 40) = 7.26, p < .01, and an interaction of Inflection with Hemisphere, F(2, 40) = 3.16, p < .05, were obtained. Resolving this interaction by hemisphere revealed for both the LH and RH effects of Inflection, F(2, 40) = 6.32–7.75, p < .005. All effects of post hoc testing seen for this latency window are displayed in Table 3, in the lower part. The results indicate the emergence of a P600 for inconsistent compared to correct items, which was broadly distributed over left and right lateral electrode sites. For consistent relative to correct items a P600 was present over right lateral electrode sites.

For midline electrode sites, an effect of Inflection was present, F(2, 40) = 9.60, p < .001, which was further analyzed by post hoc testing (see Table 3). Importantly, over midline electrode sites, the P600 for incorrect past tense items relative to their correct equivalents revealed a gradual modulation. Largest P600 amplitude was elicited by inconsistent past tense items compared to correct ones, whereas medium amplitude of P600 was elicited by consistent compared to equivalent correct past tense items.

ERPs for the Pseudoword Block

Grand average ERPs for incorrect compared to correct pseudoword items showed an N400 component emerging around 300 msec after stimulus onset. Most interestingly, the N400 was gradually modulated with largest amplitude for inconsistent pseudoword items and medium amplitude for consistent relative to correct pseudoword items (Figure 2).

Figure 2. 

Grand average ERPs of the pseudoword past tense items that were defined to be grammatically correct (blue line), consistent (green line), and inconsistent (red line). The zoom of the CP4 electrode at the bottom displays the modulation of the N400 component.

Figure 2. 

Grand average ERPs of the pseudoword past tense items that were defined to be grammatically correct (blue line), consistent (green line), and inconsistent (red line). The zoom of the CP4 electrode at the bottom displays the modulation of the N400 component.

The 300–500 msec latency window

Statistical analysis revealed a significant effect of Inflection, F(2, 40) = 13.16, p < .001, as well as significant two-way interactions of Inflection with Anterior/posterior, F(2, 40) = 8.38, p < .001, and of Inflection with Hemisphere, F(2, 40) = 5.49, p < .01. The latter interaction was resolved by hemisphere and showed significant effects of Inflection for both the LH and RH, F(2, 40) = 8.74–16.69, p < .001. This finding indicates the emergence of an N400 for incorrect relative to correct pseudoword items. The results of the post hoc testing are displayed in Table 4. The post hoc results confirm a gradual modulation of the N400 amplitude over right lateral electrode sites. Largest amplitude of N400 was obtained for inconsistent in comparison to grammatically correct pseudoword items, whereas medium amplitude of N400 was elicited by consistent in relation to correct pseudoword items.

Table 4. 

Observed Effects of the Post hoc Testing of the 300–500 msec Latency Window for the LH and RH, as well as Midline Electrode Sites

EffectLHRHMidline Sites
dftdftdft
correct vs. con 1,41 15.00*** 1,41 14.76*** 1,41 11.75*** 
correct vs. incon 1,41 15.08*** 1,41 33.72*** 1,41 22.75*** 
con vs. incon 1,41 ns 1,41 5.75** 1,41 ns 
EffectLHRHMidline Sites
dftdftdft
correct vs. con 1,41 15.00*** 1,41 14.76*** 1,41 11.75*** 
correct vs. incon 1,41 15.08*** 1,41 33.72*** 1,41 22.75*** 
con vs. incon 1,41 ns 1,41 5.75** 1,41 ns 

con = consistent; incon = inconsistent; ns = not significant.

*p < .05.

**p < .025.

***p < .016.

Statistical analysis of the midline electrode sites revealed a main effect of Inflection, F(2, 40) = 11.41, p < .001. All results of the post hoc testing are displayed in Table 4. The analysis confirms the presence of the N400 over midline electrode sites in response to incorrect relative to correct pseudoword items.

DISCUSSION

The present ERP study scrutinized the processing of irregular verb inflection with regard to morphological principles as proposed by underspecification-based accounts (Wiese, 2008). Moreover, we addressed the question whether these principles reflect rather general morphological rules by occurring independently of the lexical representation of verbs (i.e., for pseudowords). Participants read series of inflected irregular verbs embedded in syntactic contexts (e.g., er wird helfen, er hat geholfen, er half). The critical past tense form was either correct or incorrect, that is, consistent by representing a subset of expected morphosyntactic properties as for the past participle stem (e.g., *er holf) or inconsistent by missing context relevant morphosyntactic properties as for the present tense stem (e.g., *er helf). In a separate experimental block, the same types of morphological manipulations were applied to pseudoword items (e.g., er wird gechen, er hat gegochen, er gach/*goch/*gech). For the real word block, ERPs revealed a negativity-P600 pattern for incorrect compared to correct past tense forms. Most interestingly, the amplitude of P600 was gradually modulated by showing medium amplitude for consistent past tense forms and largest amplitude for inconsistent past tense forms. By contrast, for the pseudoword block a modulation of the N400 component with medium amplitude for consistent past tense forms and largest amplitude for inconsistent past tense forms relative to the correct equivalents was seen.

Findings for Irregular Real Word Verbs

In response to violated irregular inflection patterns, a negativity between 250 and 400 msec preceding a gradually modulated P600 emerged. This negativity was sensitive to the morphological anomaly of the past tense forms and showed close resemblance to the LAN component in terms of its latency and deflection. LANs were reliably seen for misapplications of morphosyntactic rules, suggesting that this brain potential is associated with morphological analysis and structure building processes (Newman et al., 2007; Morris & Holcomb, 2005; Friederici, 2002; Rodriguez-Fornells et al., 2001; Penke et al., 1997). Correspondingly, the present negativity seems to be a reflection of morphophonological analysis processes of the anomalous past tense stem forms. In contrast to most previous findings reporting a LAN, the observed negative response revealed a widespread scalp distribution, rather than showing a left anterior focus. Although previous observations of LANs were seen in response to morphological violations occurring in sentence contexts, this study applied minimal syntactic contexts (i.e., two- and three-word items). In the reduced syntactic contexts, the present morphological anomaly presumably engendered an outright violation, which might explain the large amplitude of this effect. In a study by Hasting and Kotz (2008), a similarly widespread negativity has been reported for subject–verb agreement violations employed to German two-word items (e.g., *er kegel-st vs. er kegel-t [he bowls]). Moreover, for violations of stem formation rules, widespread negativities have also been reported (Rodriguez-Fornells et al., 2001; Gross et al., 1998). These observations of more broadly distributed negativities suggest that morphological anomalies affecting the verb stem might cause more severe processing difficulty than, for instances, violations of affixation rules and apparently recruit an extended neural network.

Alternatively, the observed negativity might be speculated to resemble an N400 component as reported for violations of irregular noun inflection (Hahne et al., 2006; Lück et al., 2006; Weyerts et al., 1997). N400s were seen for regular nouns combined with an irregular affix (e.g., *Waggon-en vs. Waggon-s), which are assumed to be processed as whole forms having no lexical entry in the mental lexicon. Thus, these findings have been linked to N400 pseudoword effects. Such an interpretation, however, is rather unlikely for the present data. The present negativity was not seen for irregularized words, but for morphological anomalies within one class of verbs. Besides, at least the inconsistent past tense forms (e.g., sing) are lexically represented by comprising existing imperative forms and prevent a potential pseudoword effect. The negativity did not vary between the two incorrect past tense forms, suggesting that both forms were treated similarly during initial phases of processing. Observing a P600 component following the negativity rather suggest that the anomalous irregular verb inflection elicited a syntax-related ERP pattern. This finding of a P600 indicates that the processing of the two types of morphological anomaly might have involved reanalysis processes. In absence of any other manipulations than the morphological anomaly (e.g., probability of stimulus occurrence, task requirements), the observation of a gradual modulation of the P600 implies that, during later phases of processing, the morphosyntactic properties of the different stem forms were taken into account. Reanalysis of the consistent past tense forms sharing relevant morphosyntactic properties with the correct past tense form (i.e., the feature [+past]) appeared to be less demanding than for the inconsistent past tense forms missing those context relevant properties by being underspecified (see Wiese, 2008). Thus, the graded processing demands for the incorrect irregular past tense forms reflected by the P600 suggest that their morphosyntactic properties mattered during reanalysis of encountered parsing difficulties. The findings show that at least some irregular verbs, those described by the ABC inflection pattern, consist of subregularities, and rely on a rule-based processing as indicated by the negativity and the gradual modulation of the P600. Moreover, the ERP results are supported by the behavioral data showing that consistent past tense forms partially conformed with the relevant morphosyntactic properties of the correct equivalents were more difficult to reject than the inconsistent past tense forms. This pattern of results suggests that the processing of irregular verb inflection, when occurring in inflected verb series, not necessarily entails enhanced lexical processing, but rather resembles the processing of regular verbs by relying on morphological principles (e.g., Newman et al., 2007; Stockall & Marantz, 2006; Morris & Holcomb, 2005; Penke et al., 1997). The present data agree with morphological approaches, assuming that irregular inflection is based on systematic patterns in marking syntactic categories (e.g., Wunderlich & Fabri, 1995), instead of being unpredictable and idiosyncratic.

Findings for Irregular Pseudoword Verbs

In response to the morphological anomaly of the pseudoword items an N400 component with a typical centroparietal and right lateral scalp distribution was evoked. The obtained graded modulation of N400 between the two incorrect pseudoword past tense forms indicates that these pseudowords were processed differently because of their morphophonological characteristics. Except for the stem alternations, both pseudoword past tense forms were identical. Because neither the correct, nor the incorrect pseudowords were lexically represented in the mental lexicon, the observed N400 modulation for incorrect pseudoword items is unlikely to be associated with lexical–semantic processing (cf. Kutas & Federmeier, 2011). In line with previous studies, the present finding of N400 might rather be interpreted as prediction-related effect. For words that accord with the predicted syntactic characteristics (DeLong et al., 2005), as well as orthographic characteristics of an item (Laszlo & Federmeier, 2009), such prediction-related modulations of N400 have been observed. In the study by DeLong and colleagues (2005), for expected phonological forms of indefinite articles (i.e., a vs. an), whose syntactic characteristics matched the predicted ones, the amplitude of N400 was reduced confirming the occurrence of specific structural predictions. In this study, pseudowords were presented in a series of inflected tense forms (i.e., present tense followed by past participle and past tense), which might have induced an expectation for morphophonological characteristics of the upcoming past tense form, presumably by relying on existing irregular inflection patterns. The present findings for equivalent real irregular inflection substantiate proposed subregularities associated with observable inflection patterns. Comparable principles might have been involved in the processing of pseudoword items enabling the language comprehension system to predict particular morphophonological regularities of potential past tense forms. The processing of consistent pseudoword items comprising morphophonological characteristics of the ABB pattern appeared to be less demanding, than of inconsistent items that cannot be related to an existing inflection pattern (i.e., the ABA pattern). Note that in the pseudoword items the lexical–semantic information is reduced, allowing an examination of morphophonological information processing as such. Further evidence for an assignment of morphological principles stems from the behavioral data obtained in the appropriateness judgment. Consistent pseudoword items were more difficult than the correct and inconsistent items, suggesting that their morphophonological characteristics presumably associated with past tense interfered with the judgments. The behavioral patterns observed for the appropriateness and the grammaticality judgments corresponded with each other, which indicates that both judgments might be aligned to similar principles. Thus, the present data imply that subregularities underlying irregular verb inflection refer to a more general morphological principle, which can be incorporated in the processing of equivalent pseudoword items.

Implications for the Different Views on Regular and Irregular Verb Processing

The present finding of a syntax-related ERP pattern for irregular verbs is in accordance with underspecification-based accounts, assuming that irregular past tense inflection is based on subregularities that are processed by rule-based mechanisms (e.g., Wiese, 2008; Wunderlich & Fabri, 1995). Accordingly, stem alternations of irregular verbs represent morphosyntactic properties that have been specified by morphological principles based on a set of abstract features. The obtained behavioral and ERP data suggest that grammatically incorrect past tense forms containing a context relevant feature (i.e., [+past]) caused less processing difficulty than respective incorrect forms containing none of the relevant features (i.e., [+past, +finite]). Because both incorrect past tense forms did neither differ in their morphological, nor phonological complexity (e.g., because of differences in word length or syllable structure), the observed processing differences are apparently related to their underlying morphosyntactic properties. This processing pattern was confirmed by the present behavioral and ERP data found for the pseudoword items implying that subregularities of irregular past tense inflection refer to general morphological rules. Whereas in a previous ERP study by Opitz, Regel, Müller, and Friederici (2013) underspecification-based accounts were supported concerning the principle of compatibility shown for nominal inflection, this study provides further evidence for these accounts with regard to the principle of specificity.

The present ERP findings imply a rule-based processing of irregular verb inflection, which is inconsistent with dual-mechanism views proposing different processing mechanisms for regular and irregular verbs. Accordingly, irregular verbs are assumed to be retrieved from the mental lexicon and not to rely on rule-governed processing. The observation of a negativity and gradually modulated P600 in response to the two incorrect irregular past tense forms, however, suggest that irregular verbs seem to involve morphophonologically parsing as well as reanalysis depending on the morphosyntactic properties of the particular tense form. This ERP pattern seen for irregular verbs shows similarities with previous findings of LAN–P600 seen for overregularization of irregular verbs (see, e.g., Morris & Holcomb, 2005; Rodriguez-Fornells et al., 2001; Penke et al., 1997), as well as for regular verbs missing the respective past tense marking (Newman et al., 2007). In relation to those findings, the present data suggest similarities in the processing of regular and irregular verbs and ask for a revision of dual-mechanism models. The assumption that irregular verbs merely involve lexical–semantic processing in absence of rule-governed processing cannot be supported. In line with previous behavioral and ERP evidence on German noun inflection (Opitz et al., 2013; Penke & Krause, 2002), the current data challenge the assumptions of dual-mechanism approaches.

The present findings are clearly inconsistent with connectionist views, which propose that both inflected regular and irregular verb forms are represented as overlapping whole forms by having certain phonological and semantic features in common. The present pattern of results consisting of a negativity and modulation of the P600 in response to irregular verbs implies that an involvement of memory-based mechanisms seems rather unlikely. If the processing of irregular verbs would be driven by such mechanisms, only an N400 component, but no gradual modulation of the P600, should have been elicited by the two types of incorrect irregular past tense forms.

Taken together, the present results provide evidence for the existence of subregularities underlying irregular inflection as assumed by underspecification-based views (cf. Wiese, 2008; Wunderlich & Fabri, 1995). These subregularities seem to be independent of the lexical representation of verbs and apparently refer to a general morphological principle, as indicated by the behavioral and ERP data for the pseudoword items. By systematic manipulation of irregular past tense inflection, it was shown that at least some irregular verbs, and not only regular verbs, involve rule-based processing mechanisms.

APPENDIX A. List of Real Word Irregular German Verbs and Filler Verbs Used in Each Condition with Approximate English Translation

Present TensePast ParticiplePast Tense
CorrectConsistentInconsistent
binden (to bind) gebunden band *bund *bind 
dringen (to urge) gedrungen drang *drung *dring 
finden (to find) gefunden fand *fund *find 
gelingen (to succeed) gelungen gelang *gelung *geling 
klingen (to sound) geklungen klang *klung *kling 
mißlingen (to fail) mißlungen mißlang *mißlung *mißling 
ringen (to struggle) gerungen rang *rung *ring 
schlingen (to bolt) geschlungen schlang *schlung *schling 
schwinden (to fade) geschwunden schwand *schwund *schwind 
schwingen (to swing) geschwungen schwang *schwung *schwing 
singen (to sing) gesungen sang *sung *sing 
sinken (to sink) gesunken sank *sunk *sink 
springen (to jump) gesprungen sprang *sprung *spring 
stinken (to stink) gestunken stank *stunk *stink 
trinken (to drink) getrunken trank *trunk *trink 
winden (to wind) gewunden wand *wund *wind 
wringen (to wring) gewrungen wrang *wrung *wring 
zwingen (to force) gezwungen zwang *zwung *zwing 
bergen (to recover) geborgen barg *borg *berg 
bersten (to burst) geborsten barst *borst *berst 
brechen (to break) gebrochen brach *broch *brech 
gelten (to pertain) gegolten galt *golt *gelt 
helfen (to help) geholfen half *holf *helf 
schelten (to chide) gescholten schalt *scholt *schelt 
erschrecken (to startle) erschrocken erschrak *erschrock *erschreck 
sprechen (to speak) gesprochen sprach *sproch *sprech 
stechen (to sting) gestochen stach *stoch *stech 
sterben (to die) gestorben starb *storb *sterb 
treffen (to hit) getroffen traf *trof *treff 
verderben (to spoil) verdorben verdarb *verdorb *verderb 
werben (to advertise) geworben warb *worb *werb 
werfen (to throw) geworfen warf *worf *werf 
stehlen (to steal) gestohlen stahl *stohl *stehl 
befehlen (to order) befohlen befahl *befohl *befehl 
empfehlen (to recommend) empfohlen empfahl *empfohl *empfehl 
nehmen (to take) genommen nahm *nohm *nehm 
 
Filler Verbs 
beißen (to bite) gebissen biß – – 
gleiten (to slide) geglitten glitt – – 
kneifen (to pinch) gekniffen kniff – – 
leiden (to suffer) gelitten litt – – 
pfeifen (to whistle) gepfiffen pfiff – – 
reißen (to tear) gerissen riss – – 
reiten (to ride) geritten ritt – – 
schleichen (to creep) geschlichen schlich – – 
schleifen (to whet) geschliffen schliff – – 
schneiden (to cut) geschnitten schnitt – – 
streichen (to paint) gestrichen strich – – 
streiten (to argue) gestritten stritt – – 
Present TensePast ParticiplePast Tense
CorrectConsistentInconsistent
binden (to bind) gebunden band *bund *bind 
dringen (to urge) gedrungen drang *drung *dring 
finden (to find) gefunden fand *fund *find 
gelingen (to succeed) gelungen gelang *gelung *geling 
klingen (to sound) geklungen klang *klung *kling 
mißlingen (to fail) mißlungen mißlang *mißlung *mißling 
ringen (to struggle) gerungen rang *rung *ring 
schlingen (to bolt) geschlungen schlang *schlung *schling 
schwinden (to fade) geschwunden schwand *schwund *schwind 
schwingen (to swing) geschwungen schwang *schwung *schwing 
singen (to sing) gesungen sang *sung *sing 
sinken (to sink) gesunken sank *sunk *sink 
springen (to jump) gesprungen sprang *sprung *spring 
stinken (to stink) gestunken stank *stunk *stink 
trinken (to drink) getrunken trank *trunk *trink 
winden (to wind) gewunden wand *wund *wind 
wringen (to wring) gewrungen wrang *wrung *wring 
zwingen (to force) gezwungen zwang *zwung *zwing 
bergen (to recover) geborgen barg *borg *berg 
bersten (to burst) geborsten barst *borst *berst 
brechen (to break) gebrochen brach *broch *brech 
gelten (to pertain) gegolten galt *golt *gelt 
helfen (to help) geholfen half *holf *helf 
schelten (to chide) gescholten schalt *scholt *schelt 
erschrecken (to startle) erschrocken erschrak *erschrock *erschreck 
sprechen (to speak) gesprochen sprach *sproch *sprech 
stechen (to sting) gestochen stach *stoch *stech 
sterben (to die) gestorben starb *storb *sterb 
treffen (to hit) getroffen traf *trof *treff 
verderben (to spoil) verdorben verdarb *verdorb *verderb 
werben (to advertise) geworben warb *worb *werb 
werfen (to throw) geworfen warf *worf *werf 
stehlen (to steal) gestohlen stahl *stohl *stehl 
befehlen (to order) befohlen befahl *befohl *befehl 
empfehlen (to recommend) empfohlen empfahl *empfohl *empfehl 
nehmen (to take) genommen nahm *nohm *nehm 
 
Filler Verbs 
beißen (to bite) gebissen biß – – 
gleiten (to slide) geglitten glitt – – 
kneifen (to pinch) gekniffen kniff – – 
leiden (to suffer) gelitten litt – – 
pfeifen (to whistle) gepfiffen pfiff – – 
reißen (to tear) gerissen riss – – 
reiten (to ride) geritten ritt – – 
schleichen (to creep) geschlichen schlich – – 
schleifen (to whet) geschliffen schliff – – 
schneiden (to cut) geschnitten schnitt – – 
streichen (to paint) gestrichen strich – – 
streiten (to argue) gestritten stritt – – 

Ungrammatical past tense forms are indicated by an asterisk.

APPENDIX B. List of Pseudoword Verbs and Pseudoword Filler Verbs Used in Each Condition

Present TensePast ParticiplePast Tense
CorrectConsistentInconsistent
ginden gegunden gand *gund *gind 
mingen gemungen mang *mung *ming 
kringen gekrungen krang *krung *kring 
pinden gepunden pand *pund *pind 
gewingen gewungen gewang *gewung *gewing 
blingen geblungen blang *blung *bling 
dißtingen dißtungen dißtang *dißtung *dißting 
zingen gezungen zang *zung *zing 
schmingen geschmungen schmang *schmung *schming 
schringen geschrungen schrang *schrung *schring 
tingen getungen tang *tung *ting 
rinken gerunken rank *runk *rink 
splingen gesplungen splang *splung *spling 
stinzen gestunzen stanz *stunz *stinz 
prinken geprunken prank *prunk *prink 
schninden geschnunden schnand *schnund *schnind 
wrinden gewrunden wrand *wrund *wrind 
zwinden gezwunden zwand *zwund *zwind 
zergen gezorgen zarg *zorg *zerg 
wersten geworsten warst *worst *werst 
klechen geklochen klach *kloch *klech 
delten gedolten dalt *dolt *delt 
stelfen gestolfen stalf *stolf *stelf 
relten gerolten ralt *rolt *relt 
erschletten erschlotten erschlatt *erschlott *erschlett 
gechen gegochen gach *goch *gech 
pfechen gepfochen pfach *pfoch *pfech 
zermen gezormen zarm *zorm *zerm 
beffen geboffen baf *bof *beff 
vermerlen vermorlen vermarl *vermorl *vermerl 
werzen geworzen warz *worz *werz 
sterfen gestorfen starf *storf *sterf 
schwehen geschwohen schwah *schwoh *schweh 
berehlen berohlen berahl *berohl *berehl 
entwehmen entwommen entwahm *entwomm *entwehm 
tehnen getonnen tahn *tonn *tehn 
 
Pseudoword Filler Verbs 
feißen gefissen fiss – – 
gleifen gegliffen gliff – – 
kneiten geknitten knitt – – 
pleifen gepliffen pliff – – 
kleißen geklissen kliss – – 
geiten gegitten gitt – – 
schmeichen geschmichen schmich – – 
schleiten geschlitten schlitt – – 
wreischen gewrischen wrisch – – 
schneißen geschnissen schniss – – 
spleiten gesplitten splitt – – 
preiten gepritten pritt – – 
Present TensePast ParticiplePast Tense
CorrectConsistentInconsistent
ginden gegunden gand *gund *gind 
mingen gemungen mang *mung *ming 
kringen gekrungen krang *krung *kring 
pinden gepunden pand *pund *pind 
gewingen gewungen gewang *gewung *gewing 
blingen geblungen blang *blung *bling 
dißtingen dißtungen dißtang *dißtung *dißting 
zingen gezungen zang *zung *zing 
schmingen geschmungen schmang *schmung *schming 
schringen geschrungen schrang *schrung *schring 
tingen getungen tang *tung *ting 
rinken gerunken rank *runk *rink 
splingen gesplungen splang *splung *spling 
stinzen gestunzen stanz *stunz *stinz 
prinken geprunken prank *prunk *prink 
schninden geschnunden schnand *schnund *schnind 
wrinden gewrunden wrand *wrund *wrind 
zwinden gezwunden zwand *zwund *zwind 
zergen gezorgen zarg *zorg *zerg 
wersten geworsten warst *worst *werst 
klechen geklochen klach *kloch *klech 
delten gedolten dalt *dolt *delt 
stelfen gestolfen stalf *stolf *stelf 
relten gerolten ralt *rolt *relt 
erschletten erschlotten erschlatt *erschlott *erschlett 
gechen gegochen gach *goch *gech 
pfechen gepfochen pfach *pfoch *pfech 
zermen gezormen zarm *zorm *zerm 
beffen geboffen baf *bof *beff 
vermerlen vermorlen vermarl *vermorl *vermerl 
werzen geworzen warz *worz *werz 
sterfen gestorfen starf *storf *sterf 
schwehen geschwohen schwah *schwoh *schweh 
berehlen berohlen berahl *berohl *berehl 
entwehmen entwommen entwahm *entwomm *entwehm 
tehnen getonnen tahn *tonn *tehn 
 
Pseudoword Filler Verbs 
feißen gefissen fiss – – 
gleifen gegliffen gliff – – 
kneiten geknitten knitt – – 
pleifen gepliffen pliff – – 
kleißen geklissen kliss – – 
geiten gegitten gitt – – 
schmeichen geschmichen schmich – – 
schleiten geschlitten schlitt – – 
wreischen gewrischen wrisch – – 
schneißen geschnissen schniss – – 
spleiten gesplitten splitt – – 
preiten gepritten pritt – – 

Ungrammatical past tense forms are indicated by an asterisk.

Acknowledgments

We are grateful to Conny Schmidt and Heike Boethel for their support with the acquisition of the EEG data. Furthermore, we are thankful to all the participants for their commitment to our research. This research was supported by German Science Foundation grant (FR 519/19-1) to Gereon Müller and Angela D. Friederici.

Reprint requests should be sent to Stefanie Regel, Max-Planck-Institute for Human Cognitive and Brain Sciences, PO Box 500 355, D-04303 Leipzig, Germany, or via e-mail: regel@cbs.mpg.de.

Notes

1

Following Jacob Grimm, this class of verbs has been termed strong verbs. For convenience, however, the term “irregular verbs” is used henceforth.

2

The inflection patterns, as presented henceforth, are adopted from Wiese's (2008) analysis, showing that irregular past tense forms are differentially specific, with past tense forms being more specific than past participle forms. The composition of the patterns is based on the postulated morphological principles of the different tense forms leading to the following order: present tense–past participle–past tense. Note that this means a reordering of common inflection patterns as presented in standard grammars of German (e.g., Dudenredaktion, 2006).

3

In some languages, such as English, the pattern of AAA occurs less frequently, which describes an unaltered inflection pattern (e.g., put, let).

4

Underspecification-based approaches have been postulated mainly for nominal inflection of German by analysis of the inventory of inflectional markers (see, e.g., Blevins, 1995; Wunderlich & Fabri, 1995; Bierwisch, 1967), as well as for suppletive forms including comparatives (Bobaljik, 2012). Also see Müller (2006), Frampton (2002), Eisenberg (2000), and Wunderlich (1996) for underspecification-based approaches to verbal inflection of German, which, however, do not involve stem alternations. Other approaches that postulate systematicity of stem alternations include Segeral and Scheer (1998) and Ross (1967).

5

Similar variants of this principle are known as Subset Principle, Elsewhere Principle, Blocking Principle, Pānini's Principle, or Proper Inclusion Principle (see e.g., Stump, 2001; Halle & Marantz, 1993; Anderson, 1992, among others).

6

Note that regular and irregular verbs do not necessarily involve the same processing mechanisms. Past tense inflection of regular verbs relies on evident inflectional rules, whereas irregular past tense inflection consists of stem-internal subregularities.

7

The two incorrect conditions are termed “consistent” and “inconsistent” with regard to the occurrence of irregular inflection patterns. By consisting of past participle stems, the “consistent” condition represents an existing inflection pattern (i.e., ABB). The “inconsistent” condition, however, comprises present tense stems, thereby representing a nonexisting inflection pattern (i.e., ABA), which violates morphological principles.

8

Fp1, Fpz, Fp2, Af7, Af3, AfZ, Af4, Af8, F7, F5, F3, Fz, F4, F6, F8, Ft7, Fc5, Fc3, Fcz, Fc4, Fc6, Ft8, T7, C5, C3, Cz, C4, C6, T8, Tp7, Cp5, Cp3, Cpz, Cp4, Cp6, Tp8, P7, P5, P3, Pz, P4, P6, P8, Po7, Po3, Poz, Po4, Po8, O1, Oz, O1, and right mastoid.

REFERENCES

Allen
,
M.
,
Badecker
,
W.
, &
Osterhout
,
L.
(
2003
).
Morphological analysis in sentence processing: An ERP study
.
Language and Cognitive Processes
,
18
,
405
430
.
Anderson
,
A.
(
1992
).
Amorphous morphology
.
Cambridge, UK
:
Cambridge University Press
.
Bierwisch
,
M.
(
1967
).
Syntactic features in morphology: General problems of so-called pronominal inflection in German
. In
To honor Roman Jakobson
(pp.
239
270
).
The Hague/Paris
:
Mouton
.
Blevins
,
J. P.
(
1995
).
Syncretism and paradigmatic opposition
.
Linguistics and Philosophy
,
18
,
113
152
.
Bobaljik
,
J. D.
(
2012
).
Universals in comparativ morphology: Suppletion, superlatives, and the structure of words
.
Cambridge, MA
:
MIT Press
.
Clahsen
,
H.
(
1999
).
Lexical entries and rules of language: A multidisciplinary study of German inflection
.
Behavioral and Brain Sciences
,
22
,
991
1060
.
Coulson
,
S.
,
King
,
J. W.
, &
Kutas
,
M.
(
1998
).
Expect the unexpected: Event-related brain response to morphosyntactic violations
.
Language and Cognitive Processes
,
13
,
21
.
de Vega
,
M.
,
Urrutia
,
M.
, &
Dominguez
,
A.
(
2010
).
Tracking lexical and syntactic processes of verb morphology with ERP
.
Journal of Neurolinguistics
,
23
,
400
415
.
DeLong
,
K. A.
,
Urbach
,
T. P.
, &
Kutas
,
M.
(
2005
).
Probalistic word pre-activation during language comprehension inferred from electrical brain activity
.
Nature Neuroscience
,
8
,
1117
1121
.
Dien
,
J.
,
Spencer
,
K. M.
, &
Donchin
,
E.
(
2004
).
Parsing the late positive complex: Mental chronometry and the ERP components that inhabit the neighborhood of the P300
.
Psychophysiology
,
41
,
665
678
.
Dudenredaktion
. (
2006
).
Duden: Die deutsche Rechtschreibung
.
Mannheim, Germany
:
Dudenverlag
.
Eisenberg
,
P.
(
2000
).
Grundriss der deutschen Grammatik
(
Vol. 1: Das Wort
).
Stuttgart, Germany
:
Metzler
.
Fabricius-Hansen
,
C.
(
1977
).
Zur Klassifizierung der starken Verben im Neuhochdeutschen
.
Deutsche Sprache
,
5
,
193
205
.
Frampton
,
J.
(
2002
).
Syncretism, impoverishment, and the structure of person
. In
M.
Andronis
,
E.
Debenport
,
A.
Pycha
, &
K.
Yoshimura
(Ed.),
Papers from the Chicago Linguistics Society Meeting
(pp.
207
222
).
Chicago
:
University of Chicago
.
Friederici
,
A. D.
(
2002
).
Towards a neural basis of auditory sentence processing
.
Trends in Cognitive Sciences
,
6
,
78
84
.
Friederici
,
A. D.
, &
Frisch
,
S.
(
2000
).
Verb argument structure processing: The role of verb-specific and argument-specific information
.
Journal of Memory and Language
,
43
,
476
507
.
Friederici
,
A. D.
, &
Meyer
,
M.
(
2004
).
The brain knows the difference: Two types of grammatical violations
.
Brain Research
,
1000
,
72
77
.
Friederici
,
A. D.
,
Pfeifer
,
E.
, &
Hahne
,
A.
(
1993
).
Event-related brain potentials during natural speech processing: Effects of semantic, morphological and syntactic violations
.
Cognitive Brain Research
,
1
,
183
192
.
Gross
,
M.
,
Say
,
T.
,
Kleingers
,
M.
,
Clahsen
,
H.
, &
Münte
,
T. F.
(
1998
).
Human brain potentials to violations in morphologically complex Italian words
.
Neuroscience Letters
,
241
,
83
86
.
Gunter
,
T. C.
,
Friederici
,
A. D.
, &
Schriefers
,
H.
(
2000
).
Syntactic gender and semantic expectancy: ERPs reveal early autonomy and late interaction
.
Journal of Cognitive Neuroscience
,
12
,
556
568
.
Hagoort
,
P.
(
2003
).
Interplay between syntax and semantics during sentence comprehension: ERP effects of combining syntactic and semantic violations
.
Journal of Cognitive Neuroscience
,
15
,
883
899
.
Hagoort
,
P.
,
Brown
,
C.
, &
Groothusen
,
J.
(
1993
).
The syntactic positive shift as an ERP measure of syntactic processing
.
Language and Cognitive Processes
,
8
,
439
.
Hagoort
,
P.
,
Wassenaar
,
M.
, &
Brown
,
C. A.
(
2003
).
Syntax-related ERP-effects in Dutch
.
Cognitive Brain Research
,
16
,
38
50
.
Hahne
,
A.
,
Müller
,
J. L.
, &
Clahsen
,
H.
(
2006
).
Morphological processing in a second language: Behavioral and event-related brain potential evidence for storage and decomposition
.
Journal of Cognitive Neuroscience
,
18
,
121
134
.
Halle
,
M.
, &
Marantz
,
A.
(
1993
).
Distributed morphology and the pieces of inflection
. In
K.
Hale
&
S.
Keyser
(Eds.),
The view from building 20: Essays in linguistics in honor of Sylvain Bromberger
(pp.
111
176
).
Cambridge, MA
:
MIT Press
.
Hasting
,
A. S.
, &
Kotz
,
S. A.
(
2008
).
Speeding up syntax: On the relative timing and automaticity of local phrase structure and morphosyntactic processing as reflected in event-related brain potentials
.
Journal of Cognitive Neuroscience
,
20
,
1207
1219
.
Joanisse
,
M. F.
, &
Seidenberg
,
M. S.
(
1999
).
Impairments in verb morphology after brain injury: A connectionist model
.
Proceedings of the National Academy of Sciences, U.S.A.
,
96
,
7592
7597
.
Joanisse
,
M. F.
, &
Seidenberg
,
M. S.
(
2005
).
Imaging the past: Neural activation in frontal and temporal regions during regular and irregular past-tense processing
.
Cognitive, Affective, & Behavioral Neuroscience
,
5
,
282
296
.
Kaan
,
E.
, &
Swaab
,
T. Y.
(
2002
).
The brain circuitry of syntactic comprehension
.
Trends in Cognitive Sciences
,
6
,
350
356
.
Kutas
,
M.
, &
Federmeier
,
K. D.
(
2011
).
Thirty years and counting: Finding meaning in the N400 component of the event-related brain potential (ERP)
.
Annual Review of Psychophysiology
,
62
,
621
647
.
Kutas
,
M.
, &
Hillyard
,
S. A.
(
1983
).
Event-related brain potentials to grammatical errors and semantic anomalies
.
Memory and Cognition
,
11
,
539
550
.
Laszlo
,
S.
, &
Federmeier
,
K. D.
(
2009
).
A beautiful day in the neighborhood: An event-related potential study of lexical relationships and prediction in context
.
Journal of Memory and Language
,
61
,
326
338
.
Lehtonen
,
M.
,
Cunillera
,
T.
,
Rodríguez-Fornells
,
A.
,
Hultén
,
A.
,
Tuomainen
,
J.
, &
Laine
,
M.
(
2007
).
Recognition of morphologically complex words in Finnish: Evidence from event-related potentials
.
Brain Research
,
1148
,
123
137
.
Leinonen
,
A.
,
Brattico
,
P.
,
Jarvenpaa
,
M.
, &
Krause
,
C. M.
(
2008
).
Event-related potential (ERP) responses to violations of inflectional and derivational rules of Finnish
.
Brain Research
,
1218
,
181
193
.
Linares
,
R. E.
,
Rodriguez-Fornells
,
A.
, &
Clahsen
,
H.
(
2006
).
Stem allomorphy in the Spanish mental lexicon: Evidence from behavioral and ERP experiments
.
Brain and Language
,
97
,
110
120
.
Lück
,
M.
,
Hahne
,
A.
, &
Clahsen
,
H.
(
2006
).
Brain potentials to morphologically complex words during listening
.
Brain Research
,
1077
,
144
152
.
Mailhammer
,
R.
(
2008
).
The Germanic strong verbs. Foundations and development of a new system
.
Berlin, New York
:
Mouton de Gruyter
.
Marslen-Wilson
,
W. D.
, &
Tyler
,
L. K.
(
1998
).
Rules, representations, and the English past tense
.
Trends in Cognitive Sciences
,
2
,
428
435
.
Marslen-Wilson
,
W. D.
, &
Tyler
,
L. K.
(
2007
).
Morphology, language and the brain: The decompositional substrate for language comprehension
.
Philosophical Transactions of the Royal Society: Biological Sciences
,
362
,
823
836
.
Morris
,
J.
, &
Holcomb
,
P. J.
(
2005
).
Event-related potentials to violations of inflectional verb morphology in English
.
Cognitive Brain Research
,
25
,
963
981
.
Müller
,
G.
(
2006
).
Subanalyse verbaler Flexionsmarker
. In
E.
Breindl
,
L.
Gunkel
, &
B.
Strecker
(Ed.),
Grammatische Untersuchungen
(pp.
183
203
).
Tübingen, Germany
:
Narr
.
Münte
,
T. F.
,
Heinze
,
H. J.
, &
Mangun
,
G. R.
(
1993
).
Dissociation of brain activity related to syntactic and semantic aspects of language
.
Journal of Cognitive Neuroscience
,
5
,
335
344
.
Münte
,
T. F.
,
Matzke
,
M.
, &
Johannes
,
S.
(
1997
).
Brain activity associated with syntactic incongruencies in words and pseudo-words
.
Journal of Cognitive Neuroscience
,
9
,
318
329
.
Münte
,
T. F.
,
Say
,
T.
,
Clahsen
,
H.
,
Schiltz
,
K.
, &
Kutas
,
M.
(
1999
).
Decomposition of morphologically complex words in English: Evidence from event-related brain potentials
.
Cognitive Brain Research
,
7
,
241
253
.
Neville
,
H. J.
,
Nicol
,
J. L.
,
Barss
,
A.
,
Forster
,
K. I.
, &
Garrettdew
,
M. F.
(
1991
).
Syntactically based sentence processing classes: Evidence from event-related brain potentials
.
Journal of Cognitive Neuroscience
,
3
,
151
165
.
Newman
,
A. J.
,
Ullman
,
M. T.
,
Pancheva
,
R.
,
Waligura
,
D. L.
, &
Neville
,
H. J.
(
2007
).
An ERP study of regular and irregular English past tense inflection
.
Neuroimage
,
34
,
435
445
.
Opitz
,
A.
,
Regel
,
S.
,
Müller
,
G.
, &
Friederici
,
A. D.
(
2013
).
Neurophysiological evidence for morphological underspecification in German strong adjective inflection
.
Language
,
89
,
231
264
.
O'Rourke
,
P. L.
, &
Van Petten
,
C.
(
2011
).
Morphological agreement at a distance: Dissociation between early and late components of the event-related potential
.
Brain Research
,
1392
,
62
79
.
Osterhout
,
L.
, &
Hagoort
,
P.
(
1999
).
A superficial resemblance does not necessarily mean you are part of the family: Counterarguments to Coulson, King and Kutas (1998) in the P600/SPS-P300 debate
.
Language and Cognitive Processes
,
14
,
1
14
.
Osterhout
,
L.
, &
Holcomb
,
P. J.
(
1992
).
Event-related brain potentials elicited by syntactic anomaly
.
Journal of Memory and Language
,
31
,
785
806
.
Penke
,
M.
, &
Krause
,
M.
(
2002
).
German noun plurals: A challenge to the dual-mechanism model
.
Brain and Language
,
81
,
303
311
.
Penke
,
M.
,
Weyerts
,
H.
,
Gross
,
M.
,
Zander
,
E.
,
Münte
,
T. F.
, &
Clahsen
,
H.
(
1997
).
How the brain processes complex words: An event-related potential study of German verb inflections
.
Cognitive Brain Research
,
6
,
37
52
.
Pinker
,
S.
(
1999
).
Words and rules: The ingredients of language
.
New York
:
HarperCollins
.
Regel
,
S.
,
Gunter
,
T. C.
, &
Coulson
,
S.
(
2010
).
The communicative style of a speaker can affect language comprehension? ERP evidence from the comprehension of irony
.
Brain Research
,
1311
,
121
135
.
Regel
,
S.
,
Gunter
,
T. C.
, &
Friederici
,
A. D.
(
2011
).
Isn't it ironic? An electrophysiological exploration of figurative language processing
.
Journal of Cognitive Neuroscience
,
23
,
277
293
.
Regel
,
S.
,
Meyer
,
L.
, &
Gunter
,
T. C.
(
2014
).
Distinguishing neurocognitive processes reflected by P600 effects: Evidence from ERPs and neural oscillations
.
Plos One
,
9
,
e96840
.
Rodriguez-Fornells
,
A.
,
Clahsen
,
H.
,
Lleo
,
C.
,
Zaake
,
W.
, &
Münte
,
T. F.
(
2001
).
Event-related brain responses to morphological violations in Catalan
.
Cognitive Brain Research
,
11
,
47
58
.
Rodriguez-Fornells
,
A.
,
Münte
,
T. F.
, &
Clahsen
,
H.
(
2002
).
Morphological priming in Spanish verb forms: An ERP repetition priming study
.
Journal of Cognitive Neuroscience
,
14
,
443
454
.
Rösler
,
F.
,
Putz
,
P.
,
Friederici
,
A. F.
, &
Hahne
,
A.
(
1993
).
Event-related brain potentials while encountering semantic and syntactic constraint violations
.
Journal of Cognitive Neuroscience
,
5
,
345
362
.
Ross
,
J. R.
(
1967
).
Der Ablaut bei den deutschen starken Verben
. In
S.
G.
, VI
(Ed.),
Phonologische Studien
(
Vol. 8
, pp.
47
117
).
Berlin, Germany
:
Akademie Verlag
.
Rumelhart
,
D.
, &
McClelland
,
J. L.
(
1986
).
On learning the past tenses of English verbs. Parallel distributed processing: Explorations in the microstructure of cognition
.
Cambridge, MA
:
MIT Press
.
Schreuder
,
R.
, &
Baayen
,
R. H.
(
1995
).
Modeling morphological processing
. In
L. B.
Feldman
(Ed.),
Morphological aspects of language processing
(pp.
131
154
).
Hillsdale, NJ
:
Erlbaum
.
Segeral
,
P.
, &
Scheer
,
T.
(
1998
).
A generalized theory of Ablaut: The case of modern German strong verbs
. In
R.
Fabri
,
A.
Ortmann
, &
T.
Parodi
(Eds.),
Models of inflection
(pp.
28
59
).
Tübingen, Germany
:
Niemeyer
.
Stockall
,
L.
, &
Marantz
,
A.
(
2006
).
A single route, full decomposition model of morphological complexity: MEG evidence
.
The Mental Lexicon
,
1
,
85
123
.
Stump
,
G.
(
2001
).
Inflectional morphology
.
Cambridge
:
Cambridge University Press
.
Ullman
,
M. T.
(
2001
).
A neurocognitive perspective on language: The declarative/procedural model
.
Nature Reviews Neuroscience
,
2
,
717
726
.
Ullman
,
M. T.
(
2004
).
Contributions of memory circuits to language: The declarative/procedural model
.
Cognition
,
92
,
231
270
.
Ullman
,
M. T.
,
Corkin
,
S.
,
Coppola
,
M.
,
Hickok
,
G.
,
Growdon
,
J. H.
,
Koroshetz
,
W. J.
, et al
(
1997
).
A neural dissociation within language: Evidence the mental dictionary is part of declarative memory, and that grammatical rules are processed by the procedural system
.
Journal of Cognitive Neuroscience
,
9
,
266
276
.
Vasey
,
M. W.
, &
Thayer
,
J. F.
(
1987
).
The continuing problem of false positives in repeated measures ANOVA in psychophysiology: A multivariate solution
.
Psychophysiology
,
24
,
479
486
.
Weyerts
,
H.
,
Penke
,
M.
,
Dohrn
,
U.
,
Clahsen
,
H.
, &
Munte
,
T. F.
(
1997
).
Brain potentials indicate differences between regular and irregular German plurals
.
NeuroReport
,
8
,
957
962
.
Wiese
,
B.
(
2008
).
Form and function of verbal ablaut in contemporary standard German
. In
R.
Sackmann
(Ed.),
Explorations in integrational linguistics. Four essays on German, French, and Guaraní
(pp.
97
151
).
Amsterdam & Philadelphia
:
John Benjamins
.
Wunderlich
,
D.
(
1996
).
Minimalist morphology: The role of paradigms
. In
G.
Booij
&
J.
van Marle
(Eds.),
Yearbook of morphology 1995
(pp.
93
114
).
Dordrecht, the Netherlands
:
Kluwer
.
Wunderlich
,
D.
, &
Fabri
,
R.
(
1995
).
Minimalist morphology: An approach to inflection
.
Zeitschrift für Sprachwissenschaft
,
14
,
236
294
.