Similar but different: ERP evidence on the processing of mental and physical experiencer verbs in Malayalam

This study investigated the neurophysiological correlates of processing mental and physical subject experiencer verbs in Malayalam. Event-related brain potentials (ERPs) were recorded as 28 first-language speakers of Malayalam read intransitive sentences with the two types of experiencer verbs. Critical stimuli were either fully acceptable, whereby the subject case matched the requirements of the verb, or unacceptable, whereby the subject case violated the requirements of the verb. A linear mixed-models analysis confirmed negativity effects in the time window 400–800 ms for mental and physical experiencer verbs. Post-hoc analyses revealed that the negativity peaked relatively early for mental experiencer verbs, whereas relatively late for physical experiencer verbs. Further, the sentence-final acceptability of trials modulated the ERPs in non-anomalous conditions but not in violation conditions, and this modulation qualitatively differed between mental and physical experiencer verbs. These results suggest that, whilst a qualitatively similar mechanism is involved in the processing of both kinds of experiencer verbs, subtle but robust differences are inherent in processing mental vs. physical experiencer verbs in Malayalam.

1 Introduction

Experiencers are special, both cognitively and linguistically (Landau, 2009). Understanding and categorizing entities and activities as experiencers or non-experiencers are fundamental to how we navigate and interpret the world. This cognitive significance of experiencers has some consequences for the grammar of a language (Dabrowska, 1994; Jayaseelan, 2004; Landau, 2009; Shimoyoshi, 2015). Experiencer predicates are used across languages to convey feelings, whether they pertain to emotions such as happiness, sadness, anger, etc. (Levin and Grafmiller, 2013), or bodily sensations like pain, hunger, cold, etc. (Fleischhauer, 2016). A key feature of these verbs is that one of the participants is an experiencer (Dowty, 1991), who can be either a physical experiencer or a mental experiencer. Depending on the realization of this experiencer participant, experiencer verbs are broadly categorized into two classes: Subject Experiencer verbs (SE) and Object Experiencer verbs (OE) (Postal, 1971; Belletti and Rizzi, 1988; Lakoff, 1971; Jackendoff, 1972; Levin, 1993; Temme, 2019). Such verbs are found in various languages (1a−1d).

(1a.) The question _stim irritated Laura _exp. (English: Temme, 2019)

(1b.) Laura _exp lsoves the song _stim.

(1c.) Juan     teme los temblores(Spanish:Gonzalez,2020)

        John_exp fear the earthquakes._stim

        “John fears earthquake.”

(1d.) Los temblores        asustan a Juan.

        Theearthquakes_stim frighten Juan_exp

        “Earthquakes frighten Juan.”

These verbs differ from other verb classes because of their internal (semantic) structure, characterized by being static rather than event-oriented (Davidson, 1967; Dowty, 1979; Jackendoff, 1991; Rappaport Hovav and Levin, 2015), their argument structure and realization, which encode the verb's obligatory argument and assignment of thematic roles to the syntactic constituents (Postal, 1971; Belletti and Rizzi, 1988; Lakoff, 1971; Jackendoff, 1990; Levin, 1993).

Experiencer verbs have been an important topic of research in linguistic theories and language processing studies due to their unique syntactic (Belletti and Rizzi, 1988; Postal, 1971) and semantic (Grimshaw, 1990; Dowty, 1991; Croft, 1993; Pesetsky, 1995; Van Valin, 2005) features. Previous processing research on experiencer verbs (or psych-verbs) have broadly addressed questions about argument structure and realization, semantic complexity, and embodied cognition (for an overview, see Brennan, 2015).

Several behavioral studies have investigated the processing differences between subject experiencer verbs and object experiencer verbs in individuals with language impairments (Grodzinsky, 1995; Beretta and Campbell, 2001; Thompson and Lee, 2009, etc.) as well as in those without such impairments (Cupples, 2002; Lamers, 2007; Kretzschmar et al., 2011; Gattei et al., 2022; Mateu, 2022; Wilson and Dillon, 2022, etc.). The results of these studies showed that subject-experiencer verbs are easier to process than object-experiencer verbs, as object experiencer verbs show longer reaction time, higher error rates and significantly more fixations and regression during reading. This is because of the less complicated alignment encoding thematic role and grammatical function (Cupples, 2002; Kretzschmar et al., 2011; Gattei et al., 2015a; Gattei et al., 2017, 2022; Mateu, 2022; Wilson and Dillon, 2022) and low semantic complexities associated with subject experiencer verbs (Gong et al., 2002; Gennari and Poeppel, 2003; Frisson and Frazier, 2005; Brennan and Pylkkänen, 2008, 2010). However, studies on the subcategories of object experiencer verbs have shown that the processing of object experiencer verbs varies based on how their arguments are expressed, such as in accusative vs. dative constructions. Dative object experiencer verbs generally show higher processing costs compared to accusative ones due to differences in linking, i.e., direct vs. inverse syntax to semantic linking (Gattei et al., 2022).

Experiencer verbs have been employed in research on embodied cognition to examine how sensory-motor information contributes to semantic representation. Specifically, several neurophysiological studies have investigated the activation of motor cortices during the comprehension of concrete verbs (action verbs) and abstract verbs (experiencer verbs) (Tettamanti et al., 2005; Bedny et al., 2008, 2012; Rodríguez-Ferreiro et al., 2011; Kemmerer et al., 2013; Muraki et al., 2020, etc.). In these studies, experiencer verbs were either used as a control condition due to their lack of concrete motor and sensory associations, or they were compared to other types of abstract verbs. Findings from some of these studies indicate that various types of abstract verbs are processed differently depending on whether they pertain to emotional states, mental processes, or non-bodily states (Muraki et al., 2020). These differences are said to arise from the different representational systems (sensorimotor vs. linguistic) by each verb type.

Several ERP studies have employed experiencer verbs in investigations related to reanalysis, primarily because of the non-canonical thematic ordering inherent to such verbs and the word order preferences associated with them. These studies have mainly investigated object experiencer verbs because of their distinct reverse-linking nature and semantic complexities, with only a few studies examining the processing of subject experiencer verbs.

The studies on object experiencer verbs, many of which are on German, capitalized on the fact that dative object experiencer verbs in German are said to project a dative initial base order because the dative marked experiencer is argued to outrank the nominative case marked theme (Wunderlich, 1997, 2003; Grimshaw, 1990). (Bornkessel et al. 2002, 2003) utilized this aspect of German experiencer verbs to investigate the processing differences between object experiencer verbs and action verbs. Clause-final object experiencer verbs engendered a late positivity as opposed to active verbs, regardless of case ambiguity or word order, which was interpreted as resulting from a reanalysis occurring when the preferred thematic ordering (first argument as thematically highest) is contradicted by the verb. Similar reanalysis related late positivities have been reported in Italian (Dröge et al., 2014) and Spanish (Gattei et al., 2015b) for the processing of object experiencer verbs. (Bornkessel et al. 2004) compared object experiencer verbs with action verbs to investigate whether the lexical information associated with object experiencer verbs modulates the grammatical function reanalysis. The study revealed that clause-final object experiencer verbs engendered less pronounced negativity compared to action verbs, suggesting that their lexical information facilitates reanalysis. Additionally, a LAN was observed for object experiencer verbs in nominative initial structures compared to dative initial ones, which was said to reflect the mismatch between thematic hierarchy (Experiencer > Theme) and case hierarchy (Nominative > Dative). However, in ERP studies in which the verb is not clause-final but instead precedes the disambiguating argument, object-experiencer verbs align more closely with action verbs (Schlesewsky and Bornkessel, 2006).

Interest in the processing patterns of subject-experiencer verbs emerged after the proposition by (Bornkessel-Schlesewsky et al. 2011) that different verb types within a language might evoke distinct ERP responses to thematic reversal anomalies (henceforth TRA). This prompted studies that focused on subject experiencer verbs in the context of TRA (Bourguignon et al., 2012; Kyriaki et al., 2020). Findings from these studies suggest that subject experiencer verbs differ qualitatively from other verb types, particularly action verbs, due to differences in their thematic and aspectual structure.

The overall findings from studies on experiencer verbs can be summarized thus: (a) subject experiencer verbs are easier to process than object experiencer verbs; (b) object experiencer verbs with differences in how the experiencer argument is realized show concomitant processing differences, whereby dative experiencers incur higher processing costs than accusative experiencers.

Subject experiencer verbs express different kinds of experiences including physical experiences such as hunger, thirst, cold, etc., as in sentence (2a), and mental experiences such as happiness, sadness, etc., as in sentence (2b). However, it is not clear whether and how the processing of various types of subject experiencer verbs differs. This is because there are very few studies to date that have investigated the processing of various types of subject experiencer verbs, particularly those indicating different kinds of experiences, apart from a few exceptions, such as (Brennan and Pylkkänen 2010) and (Kretzschmar et al. 2019).

(2a.) Ich friere. (German: Klein and Kutscher, 2005)

        I be.cold

        “I am cold.”

(2b.) Tar                  khong uth.i – sil(Assamese :Subbārāo, 2012)

         He/she(GEN) anger rise.CPM-PST

         “He/she is angry.”

Whilst such verbs are present in several languages of the world, the case of Malayalam, a south Dravidian language with a canonical SOV word order, is particularly interesting in this regard because it clearly distinguishes between verbs denoting a mental experience (henceforth mental experiencer verbs) and those denoting a physical experience (henceforth physical experiencer verbs) in its syntax (Jayaseelan, 2004). Malayalam employs different case markings on the experiencer arguments based on the type of experience indicated by the experiencer verbs.

(3a.) രവിയ്ക്ക്           വിശ

         Ravi-kk                 vishannu.

         Ravi(AN.M)-DAT be. hungry-PST

         “Ravi became hungry.”

(3b.) രവി          സന്തോഷി

         Ravi             santhoshichu.

         Ravi(AN.M.NOM) be.happy – PST

          “Ravi became happy.”

As can be observed from examples (3a) and (3b), verbs pertaining to physical experiences such as “vishakkuka—be hungry” require a dative subject in the simple intransitive construction, whereas verbs associated with a mental experience like “santhoshikkuka—be happy” require a nominative subject (Mohanan, 1990; Jayaseelan, 2004; Nizar, 2010; Nair, 2012; Krishnan, 2020). The present study investigated the processing of the two types of experiencer verbs such as in (3a) and (3b) to study whether they are processed similarly or differently, given their varying morphosyntactic realization in Malayalam in spite of the fact that both are subject experiencer verbs.

The remainder of the paper is organized as follows. Section 1.1 provides a brief account of the relevance of studying experiencer verbs in Malayalam. Section 1.2 reports the motivation behind the present study and our hypothesis. Then, the material used in the study, participant information, and the procedure, the EEG recording, pre-processing, and statistical analysis are presented in the Section 2, the behavioral and ERP results are presented in Section 3. Section 4 provides a detailed discussion and interpretation of our findings, and concludes the paper.

1.1 Experiencer verbs in Malayalam

Malayalam is a nominative-accusative South Dravidian language with agglutinative morphology, and predominantly exhibits SOV word order. However, the word order is relatively flexible, with grammatical relations and semantic roles indicated in Malayalam through a series of case suffixes and/or postpositions as in other Dravidian languages. Thus, changing the word order in a sentence typically does not change its meaning, as the roles and relations are primarily conveyed through these suffixes. Malayalam verbs do not show agreement, unlike other sister languages in the Dravidian language family, and can be categorized into various types, with one significant classification being experiencer verbs or psych verbs (Jayan and Kumar, 2018). These verbs encompass semantic notions related to experiencing, wanting, feeling, liking etc. Experiencer verbs in Malayalam are further divided into two categories based on their usage within the language: simple verbs (as in 4a and 4b) and complex predicates (as in 4c and 4d) (Nizar, 2010).

(4a.) അവൾക്ക്         വിശ (Krishnan, 2020)

         Avalkk                  vishannu.

         She – DAT be.hungry – PST

         “She became hungry.”

(4b.) അവൾ     സന്തോഷി

         Aval             santhoshichu.

         She(NOM) be. – PST

         “She became happy.”

(4c.) അവൾക്ക്      വിശ       വ

         Avalkk              vishappu       vannu.

         She – DAT hunger come – PST

         “She became hungry.”

(4d.) അവൾക്ക്      സന്തോഷം     വ

         Avalkk              santhosham            vannu.

         She – DAT happiness come – PST

         “She became happy.”

As illustrated by the examples in (4a) and (4b), simple experiencer verbs in Malayalam require different cases on the subject (experiencer) argument based on the kind of experience the verb indicates (Jayaseelan, 2004). The complex predicate constructions in (4c) and (4d), on the other hand, involve predicates with two elements: the first element can be a noun, adjective, or non-finite verb, and the second element is a verb lacking semantic content known as a “light verb” (Mohanan, 1990). When these light verbs combine with a noun, adjective, or non-finite verb, the resulting complex predicate triggers the dative case marking on the logical subject. In (4c) and (4d), the first element in the complex predicate, which carries the semantic meaning of experience, is morphologically related to the simple experiencer verbs in sentences (4a) and (4b) respectively. In sum, physical experiencer verbs require the subject argument to be in the dative case, in both their simple form and corresponding complex predicate form. By contrast, mental experiencer verbs require the subject to be in the nominative case in their simple form, and in the dative case in complex predicates (Nizar, 2010; Mohanan, 1990; Krishnan, 2020; Jayaseelan, 2004).

Theoretical investigations on Mental and Physical experiencer verbs in Malayalam have shown that they differ from each other in terms of both syntax and semantics. According to Mohanan's (1990) unified account of dative-induced meanings in Malayalam using a semantic role-based approach, mental experiencer verbs denote an experience of change of state in an individual (plain experiencers), while physical experiencer verbs denote an experience of advent of a new state to an individual (dative experiencers as experiencer goals). (Nizar 2010), discussing the arguments of (Mohanan 1990), explained that the individual is the goal of the advent of a state with verbs of physical experience, and thus the subject should be marked with dative case. Because the experience of a mental experience is not a goal, it is not marked dative, but rather is in the nominative case. (Jayaseelan 2004), while discussing the subjecthood status of dative nominals, argued that mental experiencer verbs and physical experiencer verbs differ syntactically based on the presence or absence of pleonastic PRO. He also argued that the nominative subject of mental experiencer verbs indicates a greater degree of volition or control compared to the dative marked argument of physical experiencer verbs. We consider the dative marked argument of physical experiencer verbs as the subject following (Nizar 2010) and (Krishnan 2020). More importantly for present purposes, (Krishnan 2020) builds upon Butt's (2006) account of Urdu case alterations and applies it to Malayalam, a non-ergative language, and suggests that Malayalam employs the nominative vs. dative case alteration to indicate semantic differences: more control vs less control, respectively. These differences are triggered by the predicate via the theta role and the inherent/structural case assigned to the subject. (Krishnan 2020) used experiencer verbs such as “ishttapeduka—like” and “dhaahikkuka—be thirsty” cited in (Jayaseelan 2004) as examples to illustrate the presence vs. absence of volitionality/agency of a subject when it is marked nominative vs. dative respectively. When the subject is marked dative in Malayalam, it signals the fact that the state of affairs described by the predicate is not volitional, and the subject lacks control over this state. For instance, the nominative experiencer argument in examples (5a) and (5d) has relatively more control over the state, and therefore implies volition/agency, compared to the dative experiencer argument in examples (5b) and (5c).

(5a.) ഞാൻ   അവളെ   ഇഷ്ട്ടെ (Krishnan, 2020)

         Njan         avale            ishttapettu.

         I(NOM) she – ACC like – PST

         “I (willingly) liked her.”

(5b.) എനിയ്ക്ക്ക്   അവളെ   ഇഷ്ട്ടെ

         Enikk                    avale           ishttapettu.

         I – DAT she – ACC like – PST

         “I liked her (despite myself).”

(5c.) എനിയ്ക്ക്ക്   ദാഹി

         Enikk                   dhahichu.

         I – DAT thirst – PST

         “I am thirsty.”

(5d.) ഞാൻ   ദാഹി

         Njan       dhahichu.

         I(NOM) thirst – PST

         “I thirsted (for something figurative).”

However, note that examples (5b) and (5d) are exceptions rather than the rule. Mental experiencer verbs pattern together syntactically with agentive¹ verbs, and they both require nominative subjects, because the subjects of these verb types have more control/agency over the action/state. By contrast, physical experiencer verbs in their literal meaning require dative subjects. (Krishnan 2020) illustrated this with an adverb test: verbs that require nominative subjects, such as mental experiencer verbs (6a) and agentive verbs (6c) can be modified with adverbs such as “unknowingly”, “knowingly”, whereas modifying physical experiencer verbs that require dative subjects with such adverbs results in ungrammaticality (6b).

(6a.) അവൾ   അറിയാതെ   സന്തോഷി (Krishnan, 2020)

         Aval           ariyathe              santhoshichu.

         She(NOM) unknowingly be.happy – PST

         “She became happy unknowingly.

(6b.) ^* അവൾക്ക്    അറിയാതെ     വിശ

         ^*Avalkk                 ariyathe   vishannu.

         She – DAT unknowingly be.hungry – PST

         “She became hungry unknowingly.”

(6c.) അവൾ    അറിയാതെ    ർത്ഥി

         Aval            ariyathe                 prarthichu.

         She(NOM) unknowingly pray – PST

         “She prayed unknowingly.”

In view of this, the present study aimed to investigate the neurophysiological processing correlates of the two types of experiencer verbs in Malayalam.

1.2 The present study

The main goal of the current study, as mentioned in the previous sections, is to investigate whether mental and physical experiencer verbs are processed similarly, i.e., whether their neurophysiological correlates are qualitatively similar or alternatively, there are processing differences between the two.

The study employed simple intransitive subject experiencer constructions of the form Subject Noun—Experiencer Verb, because they offer unique opportunities for Dative-Nominative alteration depending on the type of experience conveyed by mental vs. physical experiencer verbs. By considering the aim of the study and structure of the experiencer verbs in Malayalam, we devised a 2 × 2 factorial design (Table 1). We manipulated the case marking of the subject argument (i.e., Nominative/Dative) and verb type, that is, mental and physical experiencer verbs. Speakers whose first-language is Malayalam read intransitive sentences comprising dative subjects (“Ravikk—Ravi(AN.M)-DAT”/“Meenuvin—Meenu(AN.F)-DAT”) or nominative subjects (“Ravi(AN.M.NOM)/Meenu(AN.F.NOM)”) followed by mental (“santhoshichu- be.happy-PST”) or physical experiencer verbs (“vishannu—be.hungry-PST”).

Table 1

Table 1. An example set of experimental stimuli.

Except for a handful of studies conducted on other Dravidian languages, including (Muralikrishnan 2007, 2011, 2024) and (Muralikrishnan et al. 2015) on Tamil, and (Niharika and Rao 2019) on Kannada, there is only one published ERP study on Malayalam that is complementary to the data reported here (Shalu et al., 2025). We have therefore considered the theoretical perspectives on mental and physical experiencer verbs in Malayalam, as discussed in the previous section, as well as the results from existing processing literature on various types of experiencer verbs while formulating the hypotheses for our study.

A straightforward hypothesis that can be derived from the cross-linguistic theoretical literature on experiencer verbs would be that there should be no differences between mental and physical experiencer verbs. This is because both verb types have the same argument structure, namely SubjectEXP–Verb_MENT/PHY (Belletti and Rizzi, 1988; Grimshaw, 1990). This would suggest that violation conditions of both verb types should elicit qualitatively similar ERP correlates compared to their correct counterparts. In terms of ERP components, we expected an N400 effect for violations involving mental as well as physical experiencer verbs (Friederici and Frisch, 2000; Dröge et al., 2014; Gattei et al., 2015b), resulting from the conflict between the lexical-semantic information of the verb and the case information of the subject noun. In line with findings from (Bornkessel et al. 2002, 2003), we further expected to observe an early positivity effect for the non-anomalous condition with mental experiencer verbs and nominative subjects, as a marker of reflecting the reanalysis from the initially preferred actor role for the nominative noun toward an experiencer role based on the verb category information. However, in view of the differences in the realization of experiencer arguments in Malayalam (Section 1.1), it is conceivable that mental and physical experiencer verbs should be processed differently in Malayalam. This would give rise to the hypothesis that we should observe neurophysiological differences between the verb types. The experiencer verbs impose different morphosyntactic constraints on their experiencer arguments in Malayalam. Mental experiencer verbs require experiencer arguments to be in the nominative case, while physical experiencer verbs require dative case marking on the experiencer argument. When these case marking requirements are not met, it can lead to processing difficulties. Additionally, we hypothesize that processing physical experiencer verbs will be more costly than processing mental experiencer verbs. There are at least three motivations for this hypothesis. First, a nominative noun in the sentence-initial position entails multiple possible continuations (Intransitive, transitive and di-transitive constructions) compared to a dative case marked noun in the sentence-initial position, with which very few unmarked or felicitous continuations are possible in the language. Second, a nominative noun would minimally predict an intransitive construction with an action verb based on the minimality principle (Bornkessel and Schlesewsky, 2006; Bornkessel-Schlesewsky and Schlesewsky, 2009), whereas a dative case marked noun would predict an experiencer verb, as cross-linguistic studies show that dative subjects are more commonly associated with experiencer verbs (Van Valin, 1991). That is, a nominative noun typically aligns with agents performing an action and thus would lead to an expectation for an action verb. When the parser encounters a physical experiencer verb following a nominative noun, this deviation from the expectation for an action verb would lead to increased processing costs. In contrast, a dative noun would predict for an experiencer verb, and when a mental experiencer verb (anomalous) is encountered instead of a physical experiencer verb, this would be less costly, because it still aligns with the more general expectation for an experiencer verb. Third, physical experiencer verbs indicate the advent of a new state whereas mental experiencer verbs indicate a change of state (Mohanan, 1990). Whilst a straightforward possibility is that there could be amplitude differences in the negativity effect for physical experiencer verbs compared to mental experiencer verbs, whether and in what manner the predicted difference between the physical and mental experiencer violation conditions is reflected in the ERPs remains to be seen.

2 Materials and participants

2.1 Ethical statement

The research protocol for this experiment was approved by the Institute Ethical Committee (IEC) of the Indian Institute of Technology Ropar, where the experiment was conducted. Informed and written consent was obtained from participants when they arrived at the lab. All the assessments thus carried out were in accordance with the guidelines and regulations approved by the committee.

2.2 Participants

Twenty-eight first-language speakers of Malayalam (mean age = 27.28; age range = 18–40; 11 female; 17 male), mostly students and staff at the Indian Institute of Technology Ropar, living in Ropar, India, participated in the experiment. Each participant was remunerated for their participation as per the allowance permitted by the ethical committee. Before the start of the experiment, the participants were briefed about the experiment and informed consent was obtained from them for the use of their data for academic purposes. All the participants were right-handed as evaluated by an adapted version of the Edinburgh Handedness Inventory (Oldfield, 1971) in Malayalam. The participants had normal or corrected-to-normal vision and had no known neurological disorder at the time of their participation in the experiment. All the participants were first-language speakers of Malayalam and reported having acquired the language before the age of six. As is the case with a sizeable part of the Indian population, most of our participants spoke additional languages. Data from 7 further participants were not included for analysis due to excessive artifacts.

2.3 Materials

We employed 9 mental experiencer verbs and 9 physical experiencer verbs for constructing the critical conditions in our study. Each verb was repeated 4 times with 4 different proper nouns (2 masculine and 2 feminine nouns) which resulted in 36 sets of sentences in four critical conditions as in Table 1, thus yielding 144 critical sentences (see Supplementary material). The subject noun was either nominative (null case marking) or dative marked. Since Malayalam dative case marker has two allomorphs “-kk” and “-”², we used an equal number of nouns requiring either the “-kk” or the “-” marker (18 each). Of the 144 critical sentences, 72 sentences were grammatically correct (half of them with dative subjects and the other half with nominative subjects) and 72 sentences were grammatically incorrect (half of them with dative subjects and the other half with nominative subjects). Additionally, we constructed filler items, which included a variety of sentence types that differed from the critical items. Specifically, they consisted of grammatical sentences with intransitive action verbs in modal construction with three different modal markers (-aam, -anam and -um), grammatical and ungrammatical complex predicate constructions (complementary study reported separately in Shalu et al., 2025), as well as grammatical transitive sentences with action verbs and dative case marked arguments. These fillers were interspersed with the critical items such that there was a total of 432 sentences, of which half were ungrammatical so that grammatical and ungrammatical sentences were equiprobable. These items were pseudorandomized for presentation during the experiment.

2.4 Procedure

The experiment consisted of a practice session followed by the actual experiment, with all the activities, including electrode preparation and stimulus presentation lasting approximately 3 h. Initially, the procedure and the tasks to be performed during the experiment were explained to the participants, and a printed instruction sheet was provided to them. However, the question under investigation in the experiment was not revealed to the participants for the acquisition of unbiased data. Then, they filled a consent form giving informed consent for their participation along with a Malayalam version of the Edinburg handedness questionnaire to evaluate their handedness. Finally, the head measurements of the participants were taken, and the Hydrocel GSN net was placed on their scalp. They were seated within a soundproof chamber on a comfortable chair at a distance of 1 m from a 20” LCD monitor on which the stimuli were presented using E-prime 2.0 (Psychology Software Tools, Pittsburgh, PA) (https://pstnet.com).

The structure of each experimental trial was as follows. Each trial commenced with a fixation “+” sign at the center of the screen for a period of 1,000 ms. After a blank screen for 100 ms, the rapid serial visual presentation of the stimulus sentence commenced, consisting of two separate chunks for critical as well as filler trials. Each of the chunks was presented for a period of 650 ms followed by an inter-stimulus interval (ISI) for a duration of 100 ms. This presentation rate provided a comfortable reading pace for participants and is in line with several existing studies on languages that are orthographically and/or morphologically complex, such as Arabic (Muralikrishnan and Idrissi, 2021), Mandarin Chinese (Wang et al., 2009), Icelandic (Bornkessel-Schlesewsky et al., 2020), Japanese (Wang and Schumacher, 2013) and Turkish (Demiral et al., 2008).

After the last chunk, participants performed two tasks. As a violation paradigm was used in the study, they performed an acceptability judgment task after the stimulus was displayed whereby, they had to press a green button on a response pad if they found the sentence displayed earlier to be acceptable or a red button if they found the sentence to be unacceptable. As a cue to the acceptability judgment task, a “???” sign appeared on the screen after the last chunk. In addition, to ensure that they read the sentence attentively, the participants performed a probe task. After finishing the acceptability judgment task or when 1,500 ms had passed, a probe word appeared on the screen for 2,000 ms. Participants were then required to press the green button if the probe word was present in the preceding sentence, or the red button if it was not. Half of the probe words were present in the preceding trial and required a “yes” response, whereas the other half were novel and required a “no” response.

The placement of the green and red buttons was counterbalanced across participants such that, the “yes” response was associated with the left key and the “no” response was associated with the right key for half the participants, and vice versa for the other half. Participants were requested to avoid blinks during the presentation of the words, but they could blink while performing the two tasks. Before the actual experiment, they were given a practice session with 12 items to familiarize themselves with the experimental trial structure. None of the actual experimental stimuli occurred in the practice session. The experimental session was subdivided into 9 blocks of 48 sentences each, with short breaks at the end of each block. At the end of the experiment, the participants were requested to fill a questionnaire regarding their experience with the experiment.

2.5 EEG recording, pre-processing and statistical analysis

The EEG was recorded by means of 32 Ag/AgCI electrodes fixed at the scalp using the Hydrocel Geodesic Sensor Net, with Cz serving as the online reference. The electrooculogram was recorded for monitoring horizontal and vertical eye movements using electrodes placed at the outer canthi and under the eyes respectively, and the interelectrode impedance was kept below 50 kΩ (amplifier input impedance > 1 GOhm) as per system recommendations (Ferree et al., 2001). All EEG and EOG signals were amplified using a Net Amps 400 Amplifier. The data was recorded at a sampling rate of 500 Hz.

The EEG data thus recorded was pre-processed for further analysis using the EEGLAB toolbox (Version 14; Delorme and Makeig, 2004, sccn.ucsd.edu) in MATLAB (Version R2023b; The MathWorks, Inc.). The data was down-sampled to 250 Hz and filtered using a 0.3–20 Hz band-pass filter in order to remove slow signal drifts. These filter settings are sufficiently broad to include language-related ERP activity that is typically in the frequency range of about 0.5–5 Hz (Delorme, 2023; Roehm et al., 2002), and have been employed in several previous cross-linguistic ERP studies on language processing in diverse languages. This data was then re-referenced offline to the average of the two mastoids, and channels on the outer extent of the face and head were removed. In a 1 Hz high-pass filtered copy of this original data, an Independent Component Analysis (ICA, Iriarte et al., 2003) was computed. Bad channels were removed in this data before submitting it to an ICA computation using the extended Infomax algorithm. The resulting Independent Components (ICs) were then tested using the ICLabel plugin (Pion-Tonachini et al., 2019) for identifying and marking ICs that were artifactual. The weights computed during ICA were then copied to the original data, and the ICs marked as artifactual were rejected from the original data. The bad channels removed before computing the ICA were interpolated from the remaining channels. The data was then imported in R (Version 4.4.2; R Core Team, 2024) using the eeguana package (Version 0.1.11.9001; Nicenboim, 2018) for epoching and statistical analysis.

Data epochs were extracted from the continuous data for each participant for the critical conditions in the critical position, i.e., the verb, from 200 ms before the onset until 1,200 ms after the onset (i.e., −200 to 1,200 ms). Epochs in which the amplitude exceeded the threshold of 100 μV in either direction were rejected, as well as those in which the difference between the minimum and maximum amplitudes within a window of 200 ms crossed the threshold of 100 μV. Furthermore, trials in which the acceptability judgement task was not performed were also rejected. Data from participants with too few remaining trials were excluded from further analysis. For a given participant, data epochs from about 27 to 28 trials remained in each condition after these rejections, and therefore the number of valid trials entering the analysis did not differ much across conditions. Across participants, there were a total of 3,228 data epochs/trials entering the analysis, with about 800 trials per critical condition. For the purposes of visualization, the valid epochs were averaged across items per condition per participant, and then the grand-averages were computed across participants and smoothed using an 8 Hz low-pass filter to obtain ERP plots at the verb for each condition.

2.5.1 ERP data analysis

The single trial EEG epochs at the NP and the verb for each critical condition were used for statistically analyzing the mean amplitudes in selected time-windows of interest by fitting linear mixed effects models using the lme4 package (Bates et al., 2015) in R (Version 4.4.2, R Core Team, 2024). The statistical models included the fixed factors Case (Nominative vs. Dative) and Verb type (Mental experiencer vs. Physical experiencer), as well as the topographical factor Regions of Interest (ROI). The ROIs were defined by clustering topographically adjacent electrodes in 6 lateral and 2 midline regions. The lateral ROIs were: Left-Frontal, comprised of the electrodes E3 and E11 (equivalent to F3 and F7 in the 10-20 electrode system); Left-Central, comprised of the electrodes E5 and E13 (C3 and T7); Left-Parietal, comprised of the electrodes E7 and E15 (P3 and P7); Right-Frontal, comprised of the electrodes E4 and E12 (F4 and F8); Right-Central, comprised of the electrodes E6 and E14 (C4 and T8); and Right-Parietal, comprised of the electrodes E8 and E16 (P4 and P8). The midline ROIs were: Mid-Fronto-Central, comprised of E17 and E28 (Fz and ~FCz); and Mid-Parieto-Occipital, comprised of E19, E20, E9, and E10 (Pz, Oz, O1, and O2). Furthermore, rather than performing a traditional subtraction-based baseline correction, the mean 200 ms pre-stimulus baseline amplitude (−200 to 0 ms) in each data epoch was included as a covariate (scaled and centered) in the statistical model. This was done in order to account for and regress out potential contributions of differences in the baseline period in the statistical analysis (Alday, 2019). However, we do not interpret effects involving prestimulus amplitudes, because these did not form part of our hypotheses. Further, this is in line with the fact that “we can include additional covariates as controls without further interpreting those covariates” (Alday, 2019, p. 9). The contrasts for the categorical factors used sum contrasts (scaled sum contrasts for 2-level factors), such that coefficients reflect differences to the grand mean (Schad et al., 2020). Except where specified otherwise, the random effects structure was maximal (Barr et al., 2013), with random intercepts for participants and items, and by-participant random slopes for the effect of Verb type, Case and their interaction term. A by-item random slope specification was not included because it would not be identifiable for this data: the items were coded uniquely across conditions such that they indicated a combination of condition + item. Following modern statistical recommendations, we do not use the term “statistically significant' or its variants based on p-value thresholds (Wasserstein et al., 2019), but report precise p-values as continuous quantities (e.g., p = 0.06 rather than p < 0.08), unless a value is “below the limit of numerical accuracy of the data”, in which case, we report it as p < 0.001 (Amrhein et al., 2019, p. 266). Further, we supplement the p-values by transforming them into s-values (Shannon information, surprisal, or binary logworth) and report s = –log₂(p), which provides a nonprobability measure of the information provided by a p-value on an absolute scale (Shannon, 1948; Greenland, 2019). In other words, “the s-value provides a gauge of the information supplied by a statistical test” and has the advantage of providing “a direct quantification of information without” requiring prior distributions as input (Rafi and Greenland, 2020, p. 6).

3 Results

3.1 Behavioral data

The mean acceptability ratings as well as the probe detection accuracy for the critical conditions, shown in Table 2, were calculated for the trials that entered the ERP analysis using the behavioral data collected during the experiment. As mentioned earlier, only those trials in which the acceptability judgement task was performed (i.e., not timed out) were considered for analysis. Acceptability was highest for the grammatical conditions, whereas it was lowest for the violation conditions. The probe detection accuracy was very high across all conditions. Figure 1 shows raincloud plots (Allen et al., 2021) of the behavioral acceptability judgements. Panel A shows the by-participant variability of acceptability ratings, with the individual data points representing the mean by-participant acceptability of each case and verb type combination. Panel B shows the by-item variability of acceptability ratings, with the individual data points representing the mean by-item acceptability of each case and verb type combination.

Table 2

Table 2. Mean acceptability ratings and probe detection accuracy.

Figure 1

Figure 1. Raincloud plot of the acceptability ratings, showing by-participant variability (A) and by-item variability (B).

The behavioral acceptability and accuracy were analyzed by fitting generalized linear mixed models using the lme4 package (Bates et al., 2015) in R. In the analysis of acceptability data, the statistical models included the fixed factors Case (Nominative vs. Dative) and Verb type (Mental experiencer vs. Physical experiencer), with random intercepts for participants and items, and by-participant random slopes for the effect of Case, Verb type and their interaction term. Type II Wald chi-square tests of the fitted model (AIC = 2352.36) of the acceptability data showed an interaction of Verb × Case [${\text{χ}}_{(1)}^2$ = 111.57, p < 0.001, s = 84.22] that was more informative than main effects of the individual factors Verb [${\text{χ}}_{(1)}^2$ = 0.08, p = 0.77, s = 0.36] and Case [${\text{χ}}_{(1)}^2$ = 5.94, p = 0.01, s = 6.08]. Estimated marginal means on the response scale were computed on the model using the emmeans package (Lenth Russell, 2021) to resolve this interaction, which showed that the estimate for mental experiencer verbs with dative subjects was less (estimate = 0.22, SE = 0.069, p = 0.001, s = 9.65) than that for mental experiencer verbs with nominative subjects (estimate = 0.97, SE = 0.009, p = 0, s = ∞). By contrast, the estimate for physical experiencer verbs with nominative subjects was less (estimate = 0.31, SE = 0.059, p < 0.001, s = 23.28) compared to that for physical experiencer verbs with dative subjects (estimate = 0.95, SE = 0.013, p = 0, s = ∞). The pairwise contrasts of these estimates within each level of Verb type showed that, the contrast between the subject types for mental experiencer verbs (estimate = 0.749, SE = 0.072, p < 0.001, s = 81.75) was similarly highly informative to that for physical experiencer verbs (estimate = −0.640, SE = 0.060, p < 0.001, s = 84.64). This is a reflection of the very low acceptability for violation conditions vs. their non-anomalous counterparts regardless of the verb type.

In the analysis of the probe detection accuracy, models with an interaction term of the fixed factors Verb and Case, as well as those with random slopes were singular. The model with Verb and Case as fixed factors with by-participant and by-item random intercepts (AIC = 846.26) detected no effects of Verb or Case (s < 1), reflecting the highly similar (ceiling) accuracy of probe detection across conditions.

3.2 ERP data

The ERPs at the critical verb position are shown in Figure 2. Visual inspection of the ERP data showed that the violation conditions engendered a negativity effect at the verb compared to their non-anomalous counterparts. In addition to considering ERP components that are relevant for language processing based on previous literature, this qualitative observation based on visual inspection informed the choice of time-window for analysis, namely 400–800 ms. This is in line with a similarly long time-window selected for analysis in a previous study on Turkish, with the critical position for analysis also at the position of the sentence-final verb (Bornkessel-Schlesewsky et al., 2011). Besides this similarity to our study, Turkish is also a verb-final language which is morphologically rich and similar (agglutinative) to Malayalam. The single trial ERP mean amplitudes extracted in the analysis time-window from a total of 3,228 data epochs/trials entered the analysis as mentioned earlier, with about a median of 800 trials per critical condition. The raw data collected during the experiment, the pre-processing pipeline used, and the pre-processed data are available in the data repository online. The ERPs at the sentence-initial noun are not part of our hypotheses but are provided for reference in Supplementary Figure S1.

Figure 2

Figure 2. Grand averaged ERPs at the verb for the critical conditions from 28 participants. Negativity is plotted upwards; the time axis runs from −0.2 to 1.2 s (i.e., −200 to 1,200 ms) with 0 being the onset of the critical verb. The blue line represents the correct mental experiencer verb (with a nominative subject), while the orange line shows the incorrect mental experiencer verb (with a dative subject), which elicited a negativity effect. The green line indicates the correct physical experiencer verb (with a dative subject), and the purple line represents the incorrect physical experiencer verb (with a nominative subject), which elicited a negativity effect.

3.3 Planned analysis

We first computed a linear mixed effects model m0 involving the fixed factors Verb type, Case of the subject noun, ROI, the covariate PrestimMeanAmp (scaled and centered), and by-participant and by-item random intercepts. The analysis code and full model outputs are available as R notebooks in the data repository online. Type II Wald Chi-squared tests on this model (AIC = 428,205.60) showed an interaction of Verb × Case [${\text{χ}}_{(1)}^2$ = 10.63, p = 0.001, s = 9.81] that was more informative than main effects of the individual factors Verb and Case (s < 1), as well as their two-way and three-way interactions with ROI (s < 1). Estimated marginal means on the response scale were computed on the model using the emmeans package (Lenth Russell, 2021) to resolve this interaction, which showed that the estimate for mental experiencer verbs with dative subjects was more negative (estimate = −1.043, SE = 0.445, p = 0.01, s = 5.70) than that for mental experiencer verbs with nominative subjects (estimate = −0.093, SE = 0.445, p = 0.83, s = 0.26). By contrast, the estimate for physical experiencer verbs with nominative subjects was more negative (estimate = −1.076, SE = 0.445, p = 0.01, s = 6) compared to that for physical experiencer verbs with dative subjects (estimate = −0.346, SE = 0.445, p = 0.43, s = 1.19). The pairwise contrasts of these estimates within each level of Verb type showed that, the contrast between the subject types for mental experiencer verbs (estimate = 0.949, SE = 0.372, p = 0.01, s = 6.52) was similar to, but slightly more informative than, that for physical experiencer verbs (estimate = −0.729, SE = 0.372, p = 0.04, s = 4.32). This pattern suggests that there is statistical support to interpret the ERPs for the violation conditions as more negative (broadly) across the scalp than their respective non-anomalous counterparts.

In order to verify if the model fit could be improved, we then computed another model m1 by making the random effects structure in m0 maximal, namely by including by-participant random slopes for the factors Verb, Case and their interaction term. As mentioned earlier, a by-item random intercept specification is not identifiable for our data in view of how items were coded in the data, and therefore the by-item random intercept specification in m0 was retained as is in m1. Such a model would better account for within-participant ERP variance between conditions. Indeed, the model fit showed an improvement (AIC = 427,161.14), and type II Wald Chi-squared tests on this improved model again showed that the interaction of Verb × Case remained more informative [${\text{χ}}_{(1)}^2$ = 3.55, p = 0.05, s = 4.07] than main effects of the individual factors Verb and Case (s < 1), as well as their two-way and three-way interactions with ROI (s < 1). Estimated marginal means on the response scale computed on model m1 to resolve this interaction showed that, as in model m0, the estimate for mental experiencer verbs with dative subjects was more negative (estimate = −1.243, SE = 0.569, p = 0.02, s = 5.11) than that for mental experiencer verbs with nominative subjects (estimate = −0.024, SE = 0.595, p = 0.96, s = 0.04). Again, as in model m0, the estimate for physical experiencer verbs with nominative subjects was more negative (estimate = −1.126, SE = 0.490, p = 0.02, s = 5.53) compared to that for physical experiencer verbs with dative subjects (estimate = −0.254, SE = 0.541, p = 0.63, s = 0.64). The pairwise contrasts of these estimates within each level of Verb type showed that, the contrast between the subject types for mental experiencer verbs (estimate = 1.219, SE = 0.778, p = 0.11, s = 3.09) was similar to, but in contrast to model m0, slightly less informative than, that for physical experiencers (estimate = −0.871, SE = 0.54, p = 0.10, s = 3.22). These numerical differences between the two models notwithstanding, the overall pattern of results remains intact, namely that there is a general effect of violation such that both violation conditions elicit larger negativity effects in this time-window than their respective non-anomalous counterparts.

There are however two unplanned observations that we made about the data that are worth noting here. One, the negativity effects for the two anomalous conditions visually appear to have slightly different morphologies. The negativity for the mental experiencer violation condition appears to show an early maximum, whereas the opposite is true for the physical experiencer violation condition. Two, although the behavioral acceptability ratings were similarly low for both violation conditions, as the raincloud plots of these ratings in Figure 1 show, a stark difference is apparent between the by-participant and by-item distribution of acceptability ratings. Whereas, the variance in by-participant ratings is similar for violations involving either verb type, the by-item distribution of the ratings is spread out for physical experiencer violations but concentrated on the lower end of the acceptability spectrum for the mental experiencer violations. In order to find out if there is some statistical support in the data for these observations, we conducted two post-hoc analyses, one each addressing each of these issues, which we report in the following section.

3.4 Post-hoc analyses

The post-hoc analyses of the data we conducted are as follows. First, in view of the slightly different morphology of the negativity effects for the two anomalous conditions, we split the original time-window for analysis reported above into two equal time-windows, namely 400–600 ms and 600–800 ms, and computed linear mixed effects models similar to m1 (i.e., with by-participant random slopes and by-item random intercepts) in each of these time-windows.³ Second, in order to investigate if the ERPs varied contingent upon the post-trial acceptability ratings, we conducted an acceptability-contingent analysis of the single trial ERP amplitudes in the original time-window (400–800 ms) for which we reported the planned comparisons above. The analysis code and full model outputs of these analyses are available as R notebooks in the data repository online.

3.4.1 Split-window analyses

Time-window 1: 400–600 ms. A linear mixed effects model similar to m1 with maximal random effects specification for the single-trial ERP amplitudes in this time-window revealed the following. Type II Wald Chi-squared tests on this model (AIC = 430,093.97) showed an interaction of Verb × Case [${\text{χ}}_{(1)}^2$ = 5.03, p = 0.02, s = 5.32] that was more informative than main effects of the individual factors Verb and Case (s ≤ 1), as well as their two-way and three-way interactions with ROI (s ≤ 1). Estimated marginal means on the response scale were computed on the model to resolve this interaction, which showed that the estimate for mental experiencer verbs with dative subjects was more negative (estimate = −1.375, SE = 0.560, p = 0.01, s = 6.14) than that for mental experiencer verbs with nominative subjects (estimate = 0.302, SE = 0.539, p = 0.57, s = 0.79). By contrast, the estimate for physical experiencer verbs with nominative subjects was more negative (estimate = −0.901, SE = 0.487, p = 0.06, s = 3.95) compared to that for physical experiencer verbs with dative subjects (estimate = −0.192, SE = 0.551, p = 0.72, s = 0.46). The pairwise contrasts of these estimates within each level of Verb type showed that, the contrast between the subject types for mental experiencer verbs (estimate = 1.678, SE = 0.731, p = 0.02, s = 5.52) was more informative than that for physical experiencer verbs (estimate = −0.708, SE = 0.567, p = 0.21, s = 2.24). This pattern suggests that the negativity effect evoked by violations is of a larger magnitude in this time-window for the mental experiencer verbs compared to that for the physical experiencer verbs.

Time-window 2: 600–800 ms. A linear mixed effects model similar to m1 with maximal random effects specification for the single-trial ERP amplitudes in this time-window revealed the following. Type II Wald Chi-squared tests on this model (AIC = 437,812.54) showed an interaction of Verb × Case [${\text{χ}}_{(1)}^2$ = 2.17, p = 0.14, s = 2.83] that was hardly more informative than main effects of the individual factors Verb and Case (s ≤ 1.5), as well as their two-way and three-way interactions with ROI (s ≤ 1). Nevertheless, estimated marginal means on the response scale were computed on the model to resolve this interaction, which showed that the estimate for mental experiencer verbs with dative subjects was more negative (estimate = −1.110, SE = 0.610, p = 0.06, s = 3.85) than that for mental experiencer verbs with nominative subjects (estimate = −0.343, SE = 0.702, p = 0.62, s = 0.67). By contrast, the estimate for physical experiencer verbs with nominative subjects was more negative (estimate = −1.352, SE = 0.537, p = 0.01, s = 6.40) compared to that for physical experiencer verbs with dative subjects (estimate = −0.310, SE = 0.579, p = 0.59, s = 0.75). The pairwise contrasts of these estimates within each level of Verb type showed that, the contrast between the subject types for mental experiencer verbs (estimate = 0.767, SE = 0.869, p = 0.37, s = 1.40) was less informative than that for physical experiencer verbs (estimate = −1.041, SE = 0.591, p = 0.07, s = 3.67). This pattern is the opposite of what was observed in the 400–600 ms time-window and suggests that the negativity effect evoked by violations is of a larger magnitude in this time-window for the physical experiencer verbs compared to that for the mental experiencer verbs.

Taken together, the split-window analyses suggest that there is statistical support to interpret the negativity effects for the violation conditions as relatively more negative for mental experiencer verbs earlier compared to physical experiencer verbs, and relatively more negative for physical experiencer verbs later compared to mental experiencer verbs. That is, whereas the negativity effect starting at about 400 ms after the onset of the verb continues to be more negative as opposed to their non-anomalous counterparts until about 800 ms for both violation conditions, there is a difference in when the negativity peaks for the two verb types. The negativity effect reaches an early maximum by 600 ms for mental experiencer verbs, whereas it reaches a later maximum after 600 ms for physical experiencer verbs.

3.4.2 Acceptability-contingent analysis

In order to examine if and in what manner the post-trial sentence acceptability possibly modulated the ERPs, we conducted an acceptability-contingent analysis. To this end, we included the sentence-final acceptability rating, coded as a factor with two levels (Acceptable and NotAcceptable; scaled sum contrasts), as a fixed effect in the linear mixed effects model m1. As mentioned earlier, the motivation for this analysis was based on the observation that there were differences apparent between the by-participant and by-item distribution of acceptability ratings for the two verb types. Consequently, we figured that the trial sequence might be crucial information to take into account when conducting this analysis. Therefore, the data epoch/trial number (scaled and centered) was included as a covariate so that possible changes in effects over the course of the experiment could be detected. As in model m1, all fixed factors and covariates and their interaction terms constituted the fixed effects in this new model m1a, and by-participant and by-item intercepts as well as by-participant random slopes for the factors Verb, Case and their interaction term constituted the random effects specification. In line with our motivations, whilst we report all detected effects, we will specifically interpret interactions of Verb type and Case that also involve the factor Acceptability in this analysis.

Model m1a (AIC = 426,765.99) proved to be a better fit of the ERP data in the 400–800 ms time-window in comparison to the models reported in the planned analysis m0 (AIC = 428,205.60) and m1 (AIC = 427,161.14). Type II Wald Chi-squared tests on model m1a showed that, the main effects of Acceptability [${\text{χ}}_{(1)}^2$ = 4.54, p = 0.03, s = 4.92] and Trial [${\text{χ}}_{(1)}^2$ = 17.26, p < 0.001, s = 14.90] and the interaction Trial × Acceptability [${\text{χ}}_{(1)}^2$ = 33.37, p < 0.001, s = 26.97] were all informative. The interactions Verb × Trial [${\text{χ}}_{(1)}^2$ = 13.01, p < 0.001, s = 11.65] and Verb × Acceptability [${\text{χ}}_{(1)}^2$ = 8.48, p = 0.003, s = 8.129] were both very informative; the interaction Case × Trial [${\text{χ}}_{(1)}^2$ = 60.82, p < 0.001, s = 47.18] was much more informative than the interaction Case × Acceptability [${\text{χ}}_{(1)}^2$ = 2.03, p = 0.153, s = 2.70]. More pertinent for present purposes, the three-way interaction Verb × Case × Trial [${\text{χ}}_{(1)}^2$ = 4.49, p = 0.03, s = 4.87] was relatively less informative than Verb × Case × Acceptability [${\text{χ}}_{(1)}^2$ = 112.01, p < 0.001, s = 84.540], which was very highly informative; the three-way interactions Verb × Trial × Acceptability [${\text{χ}}_{(1)}^2$ = 60.61, p < 0.001, s = 47.03] and Case × Trial × Acceptability [${\text{χ}}_{(1)}^2$ = 21.94, p < 0.001, s = 18.44] were both very informative. By contrast, the four-way interaction Verb × Case × Trial × Acceptability was not informative [${\text{χ}}_{(1)}^2$ = 0.30, p = 0.57, s = 0.78]. Estimated marginal means on the response scale were computed on the model to resolve the highly informative three-way interaction Verb × Case × Acceptability, which showed the following.

For mental experiencer verbs with nominative subjects (i.e., non-anomalous sentences), the estimate for sentences rated as NotAcceptable was much more negative (estimate = −4.425, SE = 0.757, p < 0.001, s = 27.55) than that for sentences rated as Acceptable (estimate = 0.153, SE = 0.602, p = 0.79, s = 0.32). By contrast, for mental experiencer verbs with dative subjects (i.e., violation sentences), the estimate for sentences rated as NotAcceptable was hardly any different (estimate = −1.288, SE = 0.552, p = 0.01, s = 5.66) from that for sentences rated as Acceptable (estimate = −1.14, SE = 0.569, p = 0.04, s = 4.47).

For physical experiencer verbs with nominative subjects (i.e., violation sentences), the estimate for sentences rated as NotAcceptable was less negative (estimate = −0.744, SE = 0.482, p = 0.12, s = 3.02) than that for sentences rated as Acceptable (estimate = −1.580, SE = 0.494, p = 0.001, s = 9.48). By contrast, for physical experiencer verbs with dative subjects (i.e., non-anomalous sentences), the estimate for sentences rated as NotAcceptable was more negative (estimate = −2.472, SE = 0.639, p = 0.0001, s = 13.15) than that for sentences rated as Acceptable (estimate = −0.125, SE = 0.541, p = 0.81, s = 0.29).

The pairwise contrasts of these estimates within each level of Verb type and Case showed that, the contrast between NotAcceptable and Acceptable ratings for mental experiencer verbs with (non-anomalous) nominative subjects (estimate = −4.579, SE = 0.476, p < 0.001, s = 70.11) was much more informative than that for mental experiencer verbs with (anomalous) dative subjects (estimate = −0.146, SE = 0.214, p = 0.49, s = 1.01). By contrast, the contrast between NotAcceptable and Acceptable for physical experiencer verbs with (anomalous) nominative subjects (estimate = 0.836, SE = 0.203, p < 0.001, s = 14.57) was relatively less informative than that for physical experiencer verbs with (non-anomalous) dative subjects (estimate = −2.346, SE = 0.365, p < 0.001, s = 32.79).

Given that the estimates are on the response scale, they are indeed estimates of the mean ERP amplitude in the 400–800 ms time-window. Therefore, the pattern above suggests that, on the one hand, there is a stark difference in the ERP amplitudes for mental experiencer verbs with nominative subjects (i.e., non-anomalous sentences) when such a sentence was rated as NotAcceptable (more negative-going ERP) vs. Acceptable. There is a similar, albeit numerically smaller, difference in the ERP amplitudes for physical experiencer verbs with dative subjects (i.e., non-anomalous sentences) when such a sentence was rated as NotAcceptable (more negative-going ERP) vs. Acceptable. On the other, there is negligible difference in the ERP amplitudes for mental experiencer verbs with dative subjects (i.e, anomalous sentences) when such a sentence was rated as NotAcceptable vs. Acceptable. Whilst there is little difference in the ERP amplitudes for physical experiencer verbs with nominative subjects (i.e., anomalous sentences) when such a sentence was rated as NotAcceptable vs. Acceptable, the ERPs for Acceptable ratings are more negative, albeit numerically small, than that for NotAcceptable ratings.

In view of the additional interaction of Verb type, Case and Trial, this intriguing pattern is best examined further by resolving the higher order interaction involving all relevant factors, namely Verb type, Case, Trial and Acceptability. Estimated marginal means on the response scale computed on the acceptability-contingent analysis model m1a to resolve this interaction are plotted in Figure 3, which illustrates the pattern described above. For each condition, the estimates (ERP amplitudes) are plotted for sentences in that condition rated as Acceptable vs. NotAcceptable. Given that the ERPs preceded the acceptability ratings in time, and were engendered during the processing, the pattern of results described above in conjunction with Figure 3 can be summarized as follows. For anomalous sentences of either verb type, there was hardly any difference between the ERPs regardless of whether the sentence was going to be rated as NotAcceptable or Acceptable. Intriguingly however, anomalous sentences with physical experiencer verbs would engender slightly more negative ERPs at the beginning of the experiment if the sentence was going to be rated as Acceptable as opposed to NotAcceptable, with this difference slowly but completely disappearing after about 1/3rd of the total number of experimental trials. That is, even though the brains of the participants detected the anomaly in both violation conditions and therefore engendered a negativity, the behavioral ratings were not always a reflection this.

Figure 3

Figure 3. Estimated marginal means for the 400–800 ms negativity time-window by Verb type, Case, Trial and sentence-final behavioral Acceptability rating. Shaded areas represent 83% confidence intervals with “mvt” multivariate adjustment for multiple comparisons. Conditions plotted in the upper-left and lower-right quadrants are grammatical; conditions plotted in the upper-right and lower-left quadrants are ungrammatical.

More interestingly however, when a non-anomalous sentence of either verb type was going to be rated as NotAcceptable as opposed to Acceptable, a concomitant negativity ensued. This negativity was consistently large over the course of the experiment for mental experiencer verbs with (non-anomalous) nominative subjects until almost the end, whereas it was almost non-existent at the beginning of the experiment for the physical experiencer verbs with (non-anomalous) dative subjects and grew larger and larger over the course of the experiment.

Simply put, the ERPs at the position of the experiencer verb are starkly different contingent upon the sentence-final acceptability for mental vs. physical experiencer verbs. Surprisingly, this difference is observed for the non-anomalous conditions rather than for the anomalous conditions.

In order to examine whether this pattern is in some way a consequence of how the acceptability ratings for the critical conditions were distributed over the course of the experiment, Figure 4 illustrates the trial-wise distribution of the acceptability ratings for the critical conditions. As can be inferred from Figure 4, the seemingly “contradictory” ratings (i.e., Acceptable ratings for ungrammatical sentences; NotAcceptable ratings for grammatical sentences) are spread over the entire course of the experiment more or less evenly rather than being clustered at a certain point in time during the experiment. So, this rules out a potential explanation for the pattern of results illustrated in Figure 3 based on possible clustering differences of ratings for the different conditions over the course of the experiment.

Figure 4

Figure 4. Trial-wise distribution of the acceptability ratings for critical sentences over the course of the experiment. Green blobs indicate trials that were rated as Acceptable. Orange blobs indicate trials that were rated as NotAcceptable. Conditions plotted in the upper-left and lower-right quadrants are grammatical; conditions plotted in the upper-right and lower-left quadrants are ungrammatical.

A tabulation of the total number of trials that entered the analysis in each condition and the percentage of those trials that were rated as either Acceptable or NotAcceptable is shown in Table 3. This data shows that a relatively small number of trials from the non-anomalous conditions were rated as NotAcceptable, whereas a considerably larger number of trials with violation sentences were rated as Acceptable. Therefore, the ERPs for the violation conditions are not very different contingent upon sentence-final acceptability, in spite of a relatively higher number of such trials that were rated as Acceptable as opposed to NotAcceptable. That is, even though there may have been enough power in the signal for violation condition trials with Acceptable ratings compared to that for violation condition trials with NotAcceptable ratings, this did not however result in a concomitant difference in the ERPs for violation conditions contingent upon sentence-final acceptability ratings. In other words, all violations resulted in very similar ERPs (negativity) regardless of whether they were rated as Acceptable or NotAcceptable. By contrast, despite the fact that relatively few non-anomalous trials with NotAcceptable ratings are evenly and similarly distributed over the entire course of the experiment for both verb types (upper-left and lower-right quadrants in Figure 4), they still resulted in very different ERP patterns for the two verb types (upper-left and lower-right quadrants in Figure 3).

Table 3

Table 3. Number and percentage of trials rated as Acceptable vs. NotAcceptable in each critical condition.

Seen in light of these details, the pattern of results in Figure 3 qualifies the ERP plots in an important way and is indeed crucial in gaining a more nuanced understanding of the processing of the two verb types. Specifically, this pattern shows that the averaged ERP plots, albeit illustrating clearly that violations correlate with a negativity effect for both verb types, do not however capture the subtle but robust difference between processing mental experiencer and physical experiencer verbs, as evidenced by very different manifestations of the negativity when non-anomalous sentences were nevertheless contradictorily judged as NotAcceptable as opposed to Acceptable. This is clearly illustrated by the plots in the upper-left and lower-right quadrants of Figure 3. Further, there is no evidence in the data for this difference to result from how the trial-wise acceptability ratings are distributed over the course of the experiment, for as Figure 4 shows, the distribution of Acceptable and NotAcceptable ratings are distributed very similarly for both conditions of a given type (i.e., on the one hand, the two violation conditions pattern together; on the other, the two non-anomalous conditions pattern together).

Consequently, seen in conjunction with:

a. the different distribution of by-item acceptability ratings for the two verb types, b. the similar trial-wise distributions of acceptability ratings for violations vs. grammatical conditions, and

c. the differences found between the violation conditions on the one hand and the non-anomalous conditions on the other as regards how the ERPs for these conditions are contingent vs. not contingent upon sentence-final acceptability ratings respectively, the stark difference apparent from Figure 3 between mental experiencer verbs with nominative subjects vs. physical experiencer verbs with dative subjects is suggestive of inherent qualitative differences in how the two types of experiencer verbs are processed in Malayalam.

3.4.3 Summary of results

One, both violation conditions elicited a negativity effect between 400 and 800 ms after the onset of the verb.

Two, post-hoc analyses to examine the slightly different morphologies of the negativity effect between the verb types showed that the negativity effect for violations with mental experiencer verbs reached an earlier maximum, whereas the negativity effect for violations with physical experiencer verbs peaked later.

Three, an acceptability-contingent analysis of the ERPs showed that the sentence-final acceptability of trials modulated the ERPs for the non-anomalous conditions but not for the violation conditions. Further, this modulation was qualitatively different between mental and physical experiencer verbs. This difference ensued despite the very similar extent and distribution of acceptability ratings for these conditions over the course of the experiment.

4 Discussion

The present study examined how mental and physical subject experiencer verbs are processed in simple intransitive constructions in Malayalam. The behavioral task showed a direct relationship between grammaticality and acceptability, with violations leading to lower acceptability compared to their grammatically correct counterparts. This pattern was consistent for both mental and physical subject experiencer verbs, despite numerical differences in the mean acceptability between the violation conditions. Electrophysiological results at the verb for violation conditions of both verb types revealed a negativity effect in the 400–800 ms time window. The post-hoc analysis carried out by splitting this time window into two confirmed that the negativity effect reached an early maximum for mental experiencer verbs (400–600 ms), whereas it reached a later maximum for physical experiencer verbs (600–800 ms). The acceptability-contingent analysis revealed that the sentence-final acceptability of trials modulated the ERPs in non-anomalous conditions but not in violation conditions. Furthermore, it showed that this modulation qualitatively differed between mental and physical experiencer verbs.

Although Figure 2 visually appears to suggest that the negativity for the physical experiencer verb violation condition was larger in amplitude compared to that for the mental experiencer verb violation condition, a closer examination of the statistical results in the analysis time window (400–800 ms) reveals that the mean amplitude for both violation conditions is, in fact, numerically almost equal: the estimate for the condition with mental experiencer verbs and dative subjects was −1.243 (SE = 0.569, p = 0.02, s = 5.11), whereas for the condition with physical experiencer verbs and nominative subjects, it was −1.126 (SE = 0.490 p = 0.02, s = 5.53). This overall pattern, i.e., negativities that are similar in amplitude for both violations, remains intact even when qualified by relative differences between the conditions in the split time-windows, in which we observed the negativity peak occurring earlier for mental experiencer verbs and later for physical experiencer verbs. Thus, in the earlier time-window in which the negativity peaked for the condition with mental experiencer verbs and dative subjects, the estimate for that condition was −1.235 (SE = 0.435, p = 0.004, s = 7.77); in the later time-window in which the negativity peaked for the condition with physical experiencer verbs and nominative subjects, the estimate for that condition was −1.337 (SE = 0.495, p = 0.006, s = 7.16). That is, the mean amplitude of the negativity was numerically similar for both verb types, with minimal difference between the two at their respective peaks.

We however did not observe the thematic reanalysis related positivity⁴ effect that we predicted for the non-anomalous condition with mental experiencer verbs following a nominative subject (nor any other late positivity effects; see Supplementary material for a language-internal motivation). This is perhaps not surprising in view of the theoretical account of experiencer verbs in Malayalam mentioned earlier (Krishnan, 2020). According to this account, mental experiencer verbs pattern together syntactically with action verbs (i.e., they both require nominative subjects), because the subjects of these verb types have more control/agency over the action/state, notwithstanding their different roles, namely experiencer vs. agent. Therefore, mental experiencer verbs, which involve a certain degree of agency compared to physical experiencer verbs, may not require thematic reanalysis. This could be attributed to the volitional nature of mental experiencer verbs, whereby the experiencer is (able to potentially) willfully engage with (or choose not to engage with) experiencing the mental state expressed by the verb. Note that a reanalysis is not relevant for physical experiencer verbs with a dative subject, because the expected thematic role of “experiencer” is already fulfilled in this context.

Our results therefore show that both mental and physical experiencer verbs in Malayalam are processed qualitatively similarly, with violations of both verb types generally evoking a negativity effect with similar latency, amplitude and topography. Nevertheless, there are subtle differences between the verb types in terms of peak latency of the effect, i.e., when the negativity effect reaches its maximum peak, as well as in the acceptability contingent analysis. In what follows, we discuss the interpretations of the negativity effect that we found for both the experiencer verbs.

In the processing literature, negativity effects that are instances of the N400 component have been interpreted in various ways. (Delogu et al. 2019) discuss three key theories about the functional interpretation of the N400 component. The access/retrieval account suggests that N400 component reflects the effort to access a word's conceptual meaning from long-term memory, with more contextual cues reducing this effort (e.g., Brouwer et al., 2017; Kutas and Federmeier, 2011). The integration account posits that N400 component indicates the difficulty of integrating a word's meaning into the existing context (e.g., Brown and Hagoort, 2000). The hybrid account combines these views, proposing that N400 reflects both retrieval and integration efforts (e.g., Baggio, 2018; Nieuwland et al., 2020). (Bornkessel-Schlesewsky and Schlesewsky 2019) provide an interpretation of all language-related negativities based on the predictive coding framework, whereby the brain continuously generates predictions about upcoming linguistic input. The amplitude modulations of negativities including the N400 are said to reflect the degree to which these predictions are confirmed or violated. Importantly, the predictions are said to be precision-weighted, with precision reflecting the relevance of a particular stimulus feature in a given language. Previous studies on experiencer verbs have reported various ERP effects at the position of the verb, including N400, P600, early positivity, and LAN. Of relevance here, the N400 effect found in the context of experiencer verbs has been mainly attributed to grammatical function reanalysis based on the preferential ordering associated with such verbs (Bornkessel et al., 2004; Schlesewsky and Bornkessel, 2006; Bornkessel-Schlesewsky et al., 2020), thematic-aspectual differences associated with them (Bourguignon et al., 2012), and violations of the verb's selection restriction (Paczynski and Kuperberg, 2011).

The negativity observed in our study could be interpreted as resulting from an integration problem, when the parser encounters a mental or physical experiencer verb following a dative or nominative subject, respectively. This occurs because the lexical-semantic information of the verb cannot be easily integrated with the experiencer subject, as the two experiencer verbs constrain the case marking of the experiencer subject argument in specific ways (Friederici and Frisch, 2000; Horiguchi, 2019). More generally, our result could also be interpreted in terms of precision-weighted prediction errors occurring at the verb due to the mismatch between top-down expectations about the verb and the bottom-up information from the verb actually encountered (Bornkessel-Schlesewsky and Schlesewsky, 2019). The prominence information of the subject noun generates predictions about the lexical-semantic features of the upcoming verb, and when these predictions are not met, it leads to processing difficulties. Evidence from previous studies that included manipulations of morphological case, word order and verb type shows that, when there is a mismatch between the previously computed prominence and the establishment of the lexical-semantic structure of the verb, this may result either in an early parietal positivity (Bornkessel et al., 2003), a late parietal positivity (Friederici and Mecklinger, 1996; Graben et al., 2000), or in a centro-parietal negativity (Bornkessel et al., 2004; Leuckefeld, 2005). In our study, both violation conditions featured sentence-final verbs that contradicted the lexical-semantic expectations about the verb and resulted in a negativity effect.

A more pertinent interpretation for our result is along the lines of (Choudhary et al. 2009) and (Nieuwland et al. 2013), in which it was posited that violations of interpretively relevant linguistic rules result in an N400 effect. In Malayalam, a clear morphosyntactic constraint governs the case marking of subject experiencers in simple experiencer constructions based on the type of experience they convey. Specifically, subject experiencer verbs denoting mental experiences require nominative subjects, while those expressing physical experiences consistently require subjects marked with the dative case (Jayaseelan, 2004). This rule is interpretively relevant because the choice of case directly influences the interpretation of the verb's meaning and reflects the structural distinction in the language. This interpretive relevance of the rule holds true equally for both verb types, as evidenced by very similar amplitudes of the negativity effect for both violation conditions.

A mental experiencer verb describes a state of change within an individual, and the individual (i.e., the experiencer argument) has more control over the state such that they can volitionally effect a change or influence it. For example, it is possible for the experiencer to willfully distract themselves from the happiness that they are experiencing, say, in order to become more balanced in their outlook or to suppress expressions of their happiness in front of others. That is, the experiencer argument has volitional control to effect a change in their mental experience without taking any tangible action. This potential for volitional control is marked by the nominative case of the mental experiencer argument. On the other hand, a physical experiencer verb describes the advent of a state, and the experiencer has hardly any control over the state and cannot effect a change in the state of affairs per se, without taking further tangible action specifically toward mitigating the physical experience. For example, a person experiencing extreme cold in freezing temperatures cannot, under normal circumstances, distract themselves to feel less cold, unless they take some tangible action (such as going indoors, wearing warm clothes etc.) toward achieving this goal. More compelling examples of this sort would be the physical experience of pain or hunger. The lack of volitional control over the state of affairs in these examples requires marking the experiencer argument with the dative case in Malayalam (Mohanan, 1990; Jayaseelan, 2004; Krishnan, 2020). In other words, marking a subject argument nominative vs. dative in Malayalam is not arbitrary; it is interpretively highly relevant, and is intended to directly convey information about the potential availability of volitional control for the experiencer argument vs. lack thereof.

As per this rule, the nominative and dative subjects in our study would have given rise to different expectations about the upcoming verb. Nominative subjects would be compatible with a range of verb types, whereas dative subjects would only be compatible with a restricted set of verb types that specifically indicate a lack of volitional control on the part of the subject argument. When the parser encounters a physical experiencer verb following a nominative subject, or a mental experiencer verb following a dative subject, as in the violation conditions in our study, the expectations generated by the subject case about the verb are not met. That is, these constitute a violation of the interpretively relevant rule that specifies the subject case based on whether or not the verb entails volitional control of the subject, which then results in engendering a negativity effect in comparison to the non-anomalous conditions. We interpret this negativity as an instance of the N400 component, in line with similar interpretations from other typologically unrelated languages such as Hindi (Choudhary et al., 2009) and Spanish (Nieuwland et al., 2013). Crucially, there was no evidence for amplitude modulations of the negativity between the two verb types in our data. This reflects the fact that the subject marking rule is equally relevant for both mental as well as physical experiencer verbs. The negativity effect found in our study showed a relatively long latency for both mental and physical experiencer verbs. Indeed, such long-latency negativities have previously been reported in several studies, both in the auditory as well as visual modalities, due to word variability across trials (Holcomb and Neville, 1990), post-lexical integration of words into higher order meaning (e.g., Brown and Hagoort, 1993; Holcomb, 1993; Steinhauer et al., 2017), complex processing required for mismatches involving verbs, which may be more demanding than those involving nouns (Courteau et al., 2019), unagreement (Ito et al., 2020), agreement violations (Bhattamishra et al., 2021), and reversal anomalies (Bornkessel-Schlesewsky et al., 2011).

Despite the fact that violations of both verb types elicited a negativity effect, there are nevertheless subtle but robust differences between the processing of the two verb types that came to light in our post-hoc analysis. A first such difference is the peak latency of the negativity effect, which occurred earlier for the mental experiencer violations but later for the physical experiencer violations. Note that this difference could be observed only when the original analysis time-window spanning from 400 to 800 ms after the onset of the verb was split into two halves, and that the overall effect is present for both verb types in the full time-window of analysis. This is in line with our hypothesis that processing violations involving physical experiencer verbs would be more costly than those with mental experiencer verbs, although we predicted that this would lead to amplitude differences (which is not the case; see above) rather than to a peak latency difference between the verb types. We speculate that a general difference in the possible range of expectations that a nominative vs. a dative subject generates could explain the observed ERP peak latency difference between the verb types. The nominative case expresses roles that have a degree of volition/control, such as prototypical agents as well as some types of experiencer arguments, whereas the dative case encodes roles that lack volition/control. Thus, as mentioned earlier, nominative subjects are compatible with a vast range of verb types, and therefore the range of possible expectations that nominative subjects generate is very broad, and in turn less specific and precise. By contrast, dative subjects in Malayalam are only compatible with a restrictive set of stative constructions, amongst which physical experiencer verbs are much more likely, as cross-linguistic studies show that dative subjects are more commonly associated with experiencer verbs (Van Valin, 1991). Consequently, the range of expectations about the verb that a dative subject generates is much more specific and precise. Seen in light of this, for violations involving mental experiencer verbs following dative subjects, the verb disconfirms a very specific expectation for the verb type, and therefore a negativity ensues and peaks relatively early on. By contrast, for violations involving physical experiencer verbs following nominative subjects, the verb disconfirms a less specific expectation. In terms of the neurobiologically plausible model of language-related negativities proposed by (Bornkessel-Schlesewsky and Schlesewsky 2019), the fact that nominative case is the default subject case for action verbs, and thus for the vast majority of verbs, adds to the decreased precision with which it can predict a specific kind of verb. Further, given that the other type of experiencer verbs (namely mental experiencer verbs) is indeed compatible with nominative subjects makes the anomaly more complex. This could potentially account for the later maximum of the negativity effect for physical experiencer verbs with a nominative subject.

This tentative explanation notwithstanding, more work is necessary in this regard. A limitation of our study is that it is not possible to disentangle the effect of differing expectations due to nominative vs. dative subjects interacting with the two types of experiencer verbs. One potential way to mitigate this limitation could involve examining mental and physical experiencer verbs within complex constructions, because the subject argument in such constructions require dative subjects for both mental and physical experiencer predicates (see examples 11 and 12). Such an investigation could help shed light on whether the observed differences in processing are driven exclusively by the verb type or not.

Nonetheless, converging evidence in support of the argument that the processing of the two verb types is indeed subtly different stems from the distinct pattern of results found for mental and physical experiencer verbs in the analysis of the ERPs contingent upon sentence acceptability. It is striking that it was the non-anomalous sentences rated as NotAcceptable that showed a clear difference between the verb types in this analysis. Indeed, the ERPs were consistently more negative for non-anomalous sentences with a mental experiencer verb following a nominative subject when they were rated as NotAcceptable as opposed to Acceptable. This difference remained prominent across trials throughout the course of the experiment until almost the end. In stark contrast, the ERPs for non-anomalous sentences with a physical experiencer verb following a dative subject showed no such difference in the beginning of the experiment, regardless of whether these were rated as NotAcceptable or Acceptable. Subsequently however, a gradual difference emerged such that the ERPs were more negative for the non-anomalous condition with physical experiencers when these were rated as NotAcceptable as opposed to Acceptable. Crucially, as the bottom right panel in Figure 3 shows, the magnitude of this difference steadily increased over the course of the experiment, becoming ever more prominent toward the end. As discussed earlier, there is no evidence in our data for an explanation for this asymmetric difference in the pattern of results between the two verb types based on potential clustering differences of ratings for the different conditions over the course of the experiment. This is because, as illustrated in Figure 4, the contradictory ratings are spread evenly across the entire experiment rather than being clustered at certain points.

The finding is particularly intriguing because the aforementioned difference in the ERPs between mental and physical experiencer verbs in non-anomalous sentences emerged from a relatively small number of trials that were rated as NotAcceptable. Furthermore, this pattern is in contrast to the ERPs for the violation conditions, which are not very different for the two verb types contingent upon sentence-final acceptability, in spite of a relatively higher number of such violation trials that were contradictorily rated as Acceptable as opposed to NotAcceptable. In short, neither the number of contradictory ratings, nor how they are spread over the course of the experiment could account for the differences observed. Instead, this pattern of results is indicative of subtle differences inherent in the processing of non-anomalous sentences of the two verb types.

In sum, the present study explored the processing of mental and physical experiencer verbs in Malayalam. The experiment revealed a negativity for subject case violations involving both mental and physical experiencer verbs, suggesting that different experiencer verb types are processed qualitatively similarly. However, post-hoc analyses of the data revealed more nuanced ERP differences between the two verb types. Specifying a more precise mechanism underlying the observed differences is outstanding, and further research is necessary to shed light on this aspect. Nevertheless, despite the fact that both mental experiencer verbs and physical experiencer verbs are processed qualitatively broadly similarly, there is sufficient evidence in our data to conclude that subtle but robust differences are inherent in processing mental vs. physical experiencer verbs in Malayalam.

Transparency statement

This article reports a complementary study to the one published as part of (Shalu et al. 2025), whereas the present study employed simple constructions to investigate the processing of experiencer verbs, the published article pertains to complex predicate constructions. That is, the two studies employed linguistically (structurally) different stimuli to study the same phenomenon from different perspectives, using alternative structures that the language under study (Malayalam) allows to convey a given meaning using very different sentence structures. The published article was prepared when the current article was already under review, and therefore it cites the pre-print of the current article. It employed structurally different sentence stimuli to confirm certain results from this study, but more importantly, to address certain issues that could not be addressed in the present study. Since both experiments were conducted together and data was collected in the same session, details pertaining to the procedure, participants and methods are identical, but otherwise the stimulus sentences that generated the data reported in the two papers as well as the analysis and findings are distinct.

Data availability statement

The statistical analysis code and full model outputs of all the analyses reported are available as R notebooks here: https://zenodo.org/records/15016472.

Ethics statement

The studies involving humans were approved by Institute Ethical Committee of the Indian Institute of Technology Ropar. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

SS: Resources, Conceptualization, Writing – original draft, Project administration, Validation, Formal analysis, Investigation, Methodology, Software, Writing – review & editing, Data curation, Visualization. RM: Supervision, Methodology, Funding acquisition, Formal analysis, Writing – review & editing, Software, Visualization, Data curation, Validation, Conceptualization. AM: Writing – review & editing, Methodology. KC: Conceptualization, Methodology, Funding acquisition, Validation, Project administration, Supervision, Writing – review & editing.

Funding

The author(s) declare that financial support was received for the research and/or publication of this article. This study was supported by the Department of Science and Technology, Government of India, through a project to Kamal Kumar Choudhary (SR/CSI/141/2012). We would like to acknowledge the Indian Institute of Technology Ropar (IIT Ropar) for the research facility maintained through an ISIRD grant to Kamal Kumar Choudhary [IITRPR/Acad./2216]. We acknowledge the Max Planck Society for covering the Article Processing Costs, which is provided through an institutional arrangement available for the corresponding author R. Muralikrishnan affiliated with the Max Planck Society.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/flang.2025.1599924/full#supplementary-material

Footnotes

1. ^Agentive verbs are prototypical action verbs that require an agent or actor argument that typically has more agent features such as volition, sentience, motion, causation etc. (Dowty, 1991). In Malayalam, prototypical agents are bear nominative case (unmarked). Whilst experiencer arguments may have some features typical of agents (such as sentience, and a degree of volition/control in some cases), they are not agents in that they do not actively initiate and cause the action whose consequence they are experiencing. Rather, they are subjected to an experience that may be physical (-control)—in which case the experiencer argument has to be marked dative in Malayalam to indicate the explicit lack of volition/control—or mental (with a degree of +control)—in which case the experiencer argument remains nominative, indicative of the fact that mental experiencers are potentially capable of a degree of volitional control over the state of affairs that they are experiencing (Mohanan, 1990).

2. ^The usage of these case suffixes is based purely on the phonology (last syllable) of the noun to which these are attached, regardless of the gender, number and other properties of the noun concerned. The allomorph “-” is used for nouns or their derivations ending in “n”, while “-kk” is applied elsewhere (Rajendran, 1977). For example, for a noun ending in “-n” in singular (such as nadan – actor(AN.M)), the appropriate dative case marker would be “-”, thus yielding “nadan” – actor-DAT. But when the same noun is in plural, as in “nadanmaar – actors”, then the correct dative case marker would be “-kk, thus yielding “nadanmaarkk- actors-DAT”, because the plural noun form does not end in “-n”. Therefore, the choice of the allomorph is purely based on the final phoneme in a noun or its derived form to which the dative case marker morpheme should be suffixed.

3. ^See Supplementary material for a data-driven motivation for these time-windows, based on the median peak latency of the negativity effects for the two violation conditions.

4. ^Whilst there is a possibility that the negativity effect may be overlapping in time to render any potential early positivity undetectable, we could not find any evidence for this in our analysis.

References

Alday, P. M. (2019). How much baseline correction do we need in ERP research? Extended GLM model can replace baseline correction while lifting its limits. Psychophysiology 56:e13451. doi: 10.1111/psyp.13451

PubMed Abstract | Crossref Full Text | Google Scholar

Allen, M., Poggiali, D., Whitaker, K., Marshall, T. R., van Langen, J., Kievit, R. A., et al. (2021). Raincloud plots: a multi-platform tool for robust data visualization. Wellcome Open Res. 4:63. doi: 10.12688/wellcomeopenres.15191.2

PubMed Abstract | Crossref Full Text | Google Scholar

Amrhein, V., Trafimow, D., and Greenland, S. (2019). Inferential statistics as descriptive statistics: there is no replication crisis if we don't expect replication. Am. Stat. 73, 262–270. doi: 10.1080/00031305.2018.1543137

Crossref Full Text | Google Scholar

Baggio, G. (2018). Meaning in the Brain. Cambridge, MA: MIT Press.

Google Scholar

Barr, D. J., Levy, R., Scheepers, C., and Tily, H. J. (2013). Random effects structure for confirmatory hypothesis testing: keep it maximal. J. Mem. Lang. 68, 255–278. doi: 10.1016/j.jml.2012.11.001

PubMed Abstract | Crossref Full Text | Google Scholar

Bates, D., Mächler, M., Bolker, B., and Walker, S. (2015). Fitting linear mixed-effects models Usinglme4. J. Stat. Softw. 67, 1–48. doi: 10.18637/jss.v067.i01

Crossref Full Text | Google Scholar

Bedny, M., Caramazza, A., Grossman, E., Pascual-Leone, A., and Saxe, R. (2008). Concepts Are more than percepts: the case of action verbs. J. Neurosci. 28, 11347–11353. doi: 10.1523/jneurosci.3039-08.2008

PubMed Abstract | Crossref Full Text | Google Scholar

Bedny, M., Caramazza, A., Pascual-Leone, A., and Saxe, R. (2012). Typical neural representations of action verbs develop without vision. Cereb. Cortex 22, 286–293. doi: 10.1093/cercor/bhr081

PubMed Abstract | Crossref Full Text | Google Scholar

Belletti, A., and Rizzi, L. (1988). Psych-verbs and θ-theory. Nat. Lang. Linguist. Theory 6, 291–352. doi: 10.1007/BF00133902

Crossref Full Text | Google Scholar

Beretta, A., and Campbell, C. (2001). Psychological verbs and the double-dependency hypothesis. Brain Cogn. 46, 42–46. doi: 10.1016/S0278-2626(01)80030-0

PubMed Abstract | Crossref Full Text | Google Scholar

Bhattamishra, S., Muralikrishnan, R., and Choudhary, K. K. (2021). Animacy modulates gender agreement comprehension in Hindi: an ERP study. Lang. Cogn. Neurosci. 37, 560–575. doi: 10.1080/23273798.2021.1980219

Crossref Full Text | Google Scholar

Bornkessel, I., McElree, B., Schlesewsky, M., and Friederici, A. D. (2004). Multi-dimensional contributions to garden path strength: dissociating phrase structure from case marking. J. Mem. Lang. 51, 495–522. doi: 10.1016/j.jml.2004.06.011

Crossref Full Text | Google Scholar

Bornkessel, I., and Schlesewsky, M. (2006). The extended argument dependency model: a neurocognitive approach to sentence comprehension across languages. Psychol. Rev. 113, 787–821. doi: 10.1037/0033-295X.113.4.787

PubMed Abstract | Crossref Full Text | Google Scholar

Bornkessel, I., Schlesewsky, M., and Friederici, A. D. (2002). Beyond syntax: language-related positivities reflect the revision of hierarchies. Neuroreport 13, 361–364. doi: 10.1097/00001756-200203040-00022

PubMed Abstract | Crossref Full Text | Google Scholar

Bornkessel, I., Schlesewsky, M., and Friederici, A. D. (2003). Eliciting thematic reanalysis effects: the role of syntax-independent information during parsing. Lang. Cogn. Processes 18, 269–298. doi: 10.1080/01690960244000018

Crossref Full Text | Google Scholar

Bornkessel-Schlesewsky, I., Kretzschmar, F., Tune, S., Wang, L., Genç, S., Philipp, M., et al. (2011). Think globally: cross-linguistic variation in electrophysiological activity during sentence comprehension. Brain Lang. 117, 133–152. doi: 10.1016/j.bandl.2010.09.010

PubMed Abstract | Crossref Full Text | Google Scholar

Bornkessel-Schlesewsky, I., Roehm, D., Mailhammer, R., and Schlesewsky, M. (2020). Language processing as a precursor to language change: evidence from Icelandic. Front. Psychol. 10:3013. doi: 10.3389/fpsyg.2019.03013

PubMed Abstract | Crossref Full Text | Google Scholar

Bornkessel-Schlesewsky, I., and Schlesewsky, M. (2009). Processing Syntax and Morphology: A Neurocognitive Perspective, Vol. 6. Oxford: Oxford University Press.

Google Scholar

Bornkessel-Schlesewsky, I., and Schlesewsky, M. (2019). Toward a neurobiologically plausible model of language-related, negative event-related potentials. Front. Psychol. 10:298. doi: 10.3389/fpsyg.2019.00298

PubMed Abstract | Crossref Full Text | Google Scholar

Bourguignon, N., Drury, J. E., Valois, D., and Steinhauer, K. (2012). Decomposing animacy reversals between agents and experiencers: an ERP study. Brain Lang. 122, 179–189. doi: 10.1016/j.bandl.2012.05.001

PubMed Abstract | Crossref Full Text | Google Scholar

Brennan, J. (2015). “Neurolinguistic processing of psychological verbs,” in International Encyclopedia of the Social andamp; Behavioral Sciences (Amsterdam: Elsevier), 670–677.

Google Scholar

Brennan, J., and Pylkkänen, L. (2008). Processing events: Behavioral and neuromagnetic correlates of Aspectual Coercion. Brain Lang. 106, 132–143. doi: 10.1016/j.bandl.2008.04.003

PubMed Abstract | Crossref Full Text | Google Scholar

Brennan, J., and Pylkkänen, L. (2010). Processing psych verbs: behavioural and MEG measures of two different types of semantic complexity. Lang. Cogn. Processes 25, 777–807. doi: 10.1080/01690961003616840

Crossref Full Text | Google Scholar

Brouwer, H., Crocker, M. W., Venhuizen, N. J., and Hoeks, J. C. (2017). A neurocomputational model of the N400 and the P600 in language processing. Cogn. Sci. 41, 1318–1352. doi: 10.1111/cogs.12461

PubMed Abstract | Crossref Full Text | Google Scholar

Brown, C., and Hagoort, P. (1993). The processing nature of the N400: evidence from masked priming. J. Cogn. Neurosci. 5, 34–44. doi: 10.1162/jocn.1993.5.1.34

PubMed Abstract | Crossref Full Text | Google Scholar

Brown, C. M., and Hagoort, P. (Eds.). (2000). The Neurocognition of Language. Oxford: Oxford University Press.

Google Scholar

Butt, M. (2006). “The dative-ergative connection,” in Empirical Issues in Syntax and Semantics, Vol. 6 (Stanford, CA: Csli Publications), 69–92.

Google Scholar

Choudhary, K. K., Schlesewsky, M., Roehm, D., and Bornkessel-Schlesewsky, I. (2009). The N400 as a correlate of interpretively relevant linguistic rules: evidence from Hindi. Neuropsychologia 473, 3012–3022. doi: 10.1016/j.neuropsychologia.2009.05.009

PubMed Abstract | Crossref Full Text | Google Scholar

Courteau, É., Martignetti, L., Royle, P., and Steinhauer, K. (2019). Eliciting ERP components for morphosyntactic agreement mismatches in perfectly grammatical sentences. Front. Psychol. 10:1152. doi: 10.3389/fpsyg.2019.01152

PubMed Abstract | Crossref Full Text | Google Scholar

Croft, W. (1993). Case marking and the semantics of mental verbs. Stud. Linguist. Philos. 49, 55–72. doi: 10.1007/978-94-011-1972-6_5

Crossref Full Text | Google Scholar

Cupples, L. (2002). The structural characteristics and on-line comprehension of experiencer-verb sentences. Lang. Cogn. Processes 17, 125–162. doi: 10.1080/01690960143000001

Crossref Full Text | Google Scholar

Dabrowska, E. (1994). Dative and nominative experiencers: two folk theories of the mind. Ling 32, 1029–1054. doi: 10.1515/ling.1994.32.6.1029

Crossref Full Text | Google Scholar

Davidson, D. (1967). The Logic of Decision and Action, Chapter The Logical Form of Action Sentences (Pittsburgh, PA: University of Pittsburgh Press), 105–148.

Google Scholar

Delogu, F., Brouwer, H., and Crocker, M. W. (2019). Event-related potentials index lexical retrieval (N400) and integration (P600) during language comprehension. Brain Cogn. 135:103569. doi: 10.1016/j.bandc.2019.05.007

PubMed Abstract | Crossref Full Text | Google Scholar

Delorme, A. (2023). EEG is better left alone. Sci. Rep. 13:2372. doi: 10.1038/s41598-023-27528-0

PubMed Abstract | Crossref Full Text | Google Scholar

Delorme, A., and Makeig, S. (2004). EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J. Neurosci. Methods 134, 9–21. doi: 10.1016/j.jneumeth.2003.10.009

PubMed Abstract | Crossref Full Text | Google Scholar

Demiral, S. B., Schlesewsky, M., and Bornkessel-Schlesewsky, I. (2008). On the universality of language comprehension strategies: evidence from Turkish. Cognition 106, 484–500. doi: 10.1016/j.cognition.2007.01.008

PubMed Abstract | Crossref Full Text | Google Scholar

Dowty, D. (1991). Thematic proto-roles and argument selection. Language 67:547. doi: 10.2307/415037

Crossref Full Text | Google Scholar

Dowty, D. R. (1979). “The semantics of aspectual classes of verbs in English,” in Word Meaning and Montague Grammar: The Semantics of Verbs and Times in Generative Semantics and in Montague's PTQ (Dordrecht: Springer Netherlands), 37–132.

Google Scholar

Dröge, A., Maffongelli, L., and Bornkessel-Schlesewsky, I. (2014). “Luigi piace a Laura?,” in Language Faculty and Beyond (Amsterdam: John Benjamins Publishing Company), 83–118.

Google Scholar

Ferree, T. C., Luu, P., Russell, G. S., and Tucker, D. M. (2001). Scalp electrode impedance, infection risk, and EEG data quality. Clin. Neurophysiol. 112, 536–544. doi: 10.1016/S1388-2457(00)00533-2

PubMed Abstract | Crossref Full Text | Google Scholar

Fleischhauer, J. (2016). Degree Gradation of Verbs, Vol. 2. Düsseldorf: Düsseldorf University Press.

Google Scholar

Friederici, A. D., and Frisch, S. (2000). Verb argument structure processing: the role of verb-specific and argument-specific information. J. Mem. Lang. 43, 476–507. doi: 10.1006/jmla.2000.2709

Crossref Full Text | Google Scholar

Friederici, A. D., and Mecklinger, A. (1996). Syntactic parsing as revealed by brain responses: first-pass and second-pass parsing processes. J. Psycholinguist. Res. 25, 157–176. doi: 10.1007/BF01708424

PubMed Abstract | Crossref Full Text | Google Scholar

Frisson, S., and Frazier, L. (2005). Carving up word meaning: portioning and grinding? J. Mem. Lang. 53, 277–291. doi: 10.1016/j.jml.2005.03.004

Crossref Full Text | Google Scholar

Gattei, C. A., Alvarez, F., París, L., Wainselboim, A., Sevilla, Y., Shalom, D., et al. (2022). Why bother? What our eyes tell about psych verb (non) causative constructions. Glossa Psycholinguistics 1:1138. doi: 10.5070/G601138

Crossref Full Text | Google Scholar

Gattei, C. A., Dickey, M. W., Wainselboim, A. J., and París, L. (2015a). The thematic hierarchy in sentence comprehension: a study on the interaction between verb class and word order in Spanish. Q. J. Exp. Psychol. 680, 1981–2007. doi: 10.1080/17470218.2014.1000345

PubMed Abstract | Crossref Full Text | Google Scholar

Gattei, C. A., Sevilla, Y., Tabullo, Á. J., Wainselboim, A. J., París, L. A., Shalom, D. E., et al. (2017). Prominence in Spanish sentence comprehension: an eye-tracking study. Lang. Cogn. Neurosci. 33, 587–607. doi: 10.1080/23273798.2017.1397278

Crossref Full Text | Google Scholar

Gattei, C. A., Tabullo, Á., París, L., and Wainselboim, A. J. (2015b). The role of prominence in Spanish sentence comprehension: an ERP study. Brain Lang. 150, 22–35. doi: 10.1016/j.bandl.2015.08.001

PubMed Abstract | Crossref Full Text | Google Scholar

Gennari, S., and Poeppel, D. (2003). Processing correlates of lexical semantic complexity. Cognition 89, B27–B41. doi: 10.1016/S0010-0277(03)00069-6

PubMed Abstract | Crossref Full Text | Google Scholar

Gong, Y., McKoon, G., and Macfarland, T. (2002). Event templates in the lexical representations of verbs. Cogn. Psychol. 45, 1–44. doi: 10.1016/S0010-0285(02)00004-X

PubMed Abstract | Crossref Full Text | Google Scholar

Gonzalez, B. H. (2020). The syntactic distribution of object experiencer psych verbs in Heritage Spanish. Languages 5:63. doi: 10.3390/languages5040063

Crossref Full Text | Google Scholar

Graben, P. B., Saddy, J. D., Schlesewsky, M., and Kurths, J. (2000). Symbolic dynamics of event-related brain potentials. Phys. Rev. E 62, 5518–5541. doi: 10.1103/PhysRevE.62.5518

PubMed Abstract | Crossref Full Text | Google Scholar

Greenland, S. (2019). Valid P-values behave exactly as they should: some misleading criticisms of P-values and their resolution with S-values. Am. Stat. 73, 106–114. doi: 10.1080/00031305.2018.1529625

Crossref Full Text | Google Scholar

Grimshaw, J. (1990). Argument Structure. Cambridge, Mass: MIT Press.

Google Scholar

Grodzinsky, Y. (1995). A restrictive theory of agrammatic comprehension. Brain Lang. 50, 27–51. doi: 10.1006/brln.1995.1039

PubMed Abstract | Crossref Full Text | Google Scholar

Holcomb, P. J. (1993). Semantic priming and stimulus degradation: implications for the role of the N400 in language processing. Psychophysiology 30, 47–61. doi: 10.1111/j.1469-8986.1993.tb03204.x

PubMed Abstract | Crossref Full Text | Google Scholar

Holcomb, P. J., and Neville, H. J. (1990). Auditory and visual semantic priming in lexical decision: a comparison using event-related brain potentials. Lang. Cogn. Processes 5, 281–312. doi: 10.1080/01690969008407065

Crossref Full Text | Google Scholar

Horiguchi, Y. (2019). The Implicit Checking of Lexical Relation Between Verbs and Objects in the Processing of Case Information: An ERP Study (Unpublished master's thesis). Tokyo Metropolitan University, Tokyo, Japan.

Google Scholar

Iriarte, J., Urrestarazu, E., Valencia, M., Alegre, M., Malanda, A., Viteri, C., et al. (2003). Independent component analysis as a tool to eliminate artifacts in EEG: a quantitative study. J. Clin. Neurophysiol. 20, 249–257. doi: 10.1097/00004691-200307000-00004

PubMed Abstract | Crossref Full Text | Google Scholar

Ito, A., Gambi, C., Pickering, M. J., Fuellenbach, K., and Husband, E. M. (2020). Prediction of phonological and gender information: an event-related potential study in Italian. Neuropsychologia 136:107291. doi: 10.1016/j.neuropsychologia.2019.107291

PubMed Abstract | Crossref Full Text | Google Scholar

Jackendoff, R. (1990). On Larson's treatment of the double object construction. Linguist. Inquiry 21:427–456.

Google Scholar

Jackendoff, R. (1991). Parts and boundaries. Cognition 41, 9–45. doi: 10.1016/0010-0277(91)90031-X

PubMed Abstract | Crossref Full Text | Google Scholar

Jackendoff, R. S. (1972). Semantic interpretation in generative grammar. Cambridge, MA: MIT Press.

Google Scholar

Jayan, J. P., and Kumar, J. S. (2018). Levin's classification on Malayalam verbs. Int. J. Linguist. Literat. (IJLL), 7, 61–72.

Google Scholar

Jayaseelan, K. A. (2004). “The possessor—experiencer dative in Malayalam,” in Typological Studies in Language (Amsterdam: John Benjamins Publishing Company), 227–244.

Google Scholar

Kemmerer, D., Miller, L., MacPherson, M. K., Huber, J., and Tranel, D. (2013). An investigation of semantic similarity judgments about action and non-action verbs in Parkinson's disease: implications for the embodied cognition framework. Front. Human Neurosci. 7:146. doi: 10.3389/fnhum.2013.00146

PubMed Abstract | Crossref Full Text | Google Scholar

Klein, K., and Kutscher, S. (2005). Lexical Economy and Case Selection of Psych-Verbs in German. Available online at: http://www.linguistics.ruhr-uni-bochum.de (Accessed November 28, 2025).

Google Scholar

Kretzschmar, F., Bornkessel-Schlesewsky, I., Staub, A., Roehm, D., and Schlesewsky, M. (2011). “Prominence facilitates ambiguity resolution: on the interaction between referentiality, thematic roles and word order in syntactic reanalysis,” in Studies in Theoretical Psycholinguistics, Vol. 36 (Dordrecht: Springer), 239–271.

Google Scholar

Kretzschmar, F., Graf, T., Philipp, M., and Primus, B. (2019). An Experimental Investigation of Agent Prototypicality and Agent Prominence in German (Tübingen: Universitätsbibliothek Tübingen), 101–123.

Google Scholar

Krishnan, G. (2020). Malayalam Morphosyntax: Inflectional Features and their Acquisition. Charlottesville, VA: Center for Open Science.

Google Scholar

Kutas, M., and Federmeier, K. D. (2011). Thirty years and counting: finding meaning in the N400 component of the event-related brain potential (ERP). Ann. Rev. Psychol. 62, 621–647. doi: 10.1146/annurev.psych.093008.131123

PubMed Abstract | Crossref Full Text | Google Scholar

Kyriaki, L., Schlesewsky, M., and Bornkessel-Schlesewsky, I. (2020). Semantic reversal anomalies under the microscope: task and modality influences on language-associated event-related potentials. Eur. J. Neurosci. 52, 3803–3827. doi: 10.1111/ejn.14862

PubMed Abstract | Crossref Full Text | Google Scholar

Lakoff, G. (1971). Irregularity in Syntax, Holt, Rinehart and Winston. New York: Holt, Rinehart and Winston.

Google Scholar

Lamers, M. J. (2007). Verb type, animacy and definiteness in grammatical function disambiguation. Linguist. Netherlands 24:125. doi: 10.1075/avt.24.13lam

Crossref Full Text | Google Scholar

Landau, I. (2009). The Locative Syntax of Experiencers. Cambridge, MA: MIT Press.

Google Scholar

Lenth Russell, V. (2021). emmeans: Estimated Marginal Means, aka Least-Squares Means. R package version 1.6.0. Available online at: https://CRAN.R-project.org/package=emmeans (Accessed Novemebr 28, 2025).

Google Scholar

Leuckefeld, K. (2005). The development of argument processing mechanisms in German: An electrophysiological investigation with school-aged children and adults (Doctoral dissertation). Max Planck Institute for Human Cognitive and Brain Sciences Leipzig, Leipzig.

Google Scholar

Levin, B. (1993). English Verb Classes and Alternations: A Preliminary Investigation. Chicago, IL: University of Chicago Press.

Google Scholar

Levin, B., and Grafmiller, J. (2013). Do You Always Fear What Frightens You? Stanford, CA: CSLI Publications.

Google Scholar

Mateu, V. (2022). Object wh-questions with psych verbs are easy in child Spanish. Glossa J. General Linguist. 7, 1–28. doi: 10.16995/glossa.7902

Crossref Full Text | Google Scholar

Mohanan, T. (1990). “Dative subjects in Malayalam: semantic information in syntax,” in Experiencer Subjects in South Asian Languages, ed. K. P. Mohanan (Stanford, CA: Stanford University), 43–66.

Google Scholar

Muraki, E. J., Cortese, F., Protzner, A. B., and Pexman, P. M. (2020). Heterogeneity in abstract verbs: an ERP study. Brain Lang. 211:104863. doi: 10.1016/j.bandl.2020.104863

PubMed Abstract | Crossref Full Text | Google Scholar

Muraki, E. J., Cortese, F., Protzner, A. B., and Pexman, P. M. (2020). Heterogeneity in abstract verbs: an ERP study. Brain Lang. 211:104863. doi: 10.1016/j.bandl.2020.104863

PubMed Abstract | Crossref Full Text | Google Scholar

Muralikrishnan, R. (2007). The influence of Word-order and Animacy in processing transitive sentences: Neurophysiological evidence from Tamil. Zenodo (Unpublished MA thesis). Saarland University, Saarland University.

Google Scholar

Muralikrishnan, R. (2011). An electrophysiological investigation of Tamil dative-subject constructions (Doctoral dissertation). Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig.

Google Scholar

Muralikrishnan, R. (2024). “Electrophysiological Investigations of Sentence Processing in Tamil,” in The Oxford Handbook of Dravidian Languages. Oxford: Oxford University Press.

Google Scholar

Muralikrishnan, R., and Idrissi, A. (2021). Salience-weighted agreement feature hierarchy modulates language comprehension. Cortex 141, 168–189. doi: 10.1016/j.cortex.2021.03.029

PubMed Abstract | Crossref Full Text | Google Scholar

Muralikrishnan, R., Schlesewsky, M., and Bornkessel-Schlesewsky, I. (2015). Animacy-based predictions in language comprehension are robust: contextual cues modulate but do not nullify them. Brain Res. 1608, 108–137. doi: 10.1016/j.brainres.2014.11.046

PubMed Abstract | Crossref Full Text | Google Scholar

Nair, R. S. S. (2012). Semantics of the Dative Case in Malayalam. Lang. India, 12, 1–15.

Google Scholar

Nicenboim, B. (2018). eeguana: A package for manipulating EEG data in R. Computer software Version 0.1.11.9001. Available online at: https://github.com/bnicenboim/eeguana (Accessed November 28, 2025).

Google Scholar

Nieuwland, M. S., Barr, D. J., Bartolozzi, F., Busch-Moreno, S., Darley, E., Donaldson, D., et al. (2020). Dissociable effects of prediction and integration during language comprehension: evidence from a large-scale study using brain potentials. Philos. Transac. Royal Soc. B. 375:20180522. doi: 10.1098/rstb.2018.0522

PubMed Abstract | Crossref Full Text | Google Scholar

Nieuwland, M. S., Martin, A. E., and Carreiras, M. (2013). Event-related brain potential evidence for animacy processing asymmetries during sentence comprehension. Brain Lang. 126, 151–158. doi: 10.1016/j.bandl.2013.04.005

PubMed Abstract | Crossref Full Text | Google Scholar

Niharika, M. K., and Rao, P. K. S. (2019). Behavioral and electrophysiological correlates of semantic processing in Kannada. Res. Rev. J. Neurosci. 7, 15–25.

Google Scholar

Nizar, M. (2010). Dative Subject Constructions in South-Dravidian Languages (Unpublished master's thesis). University of California, Berkely.

Google Scholar

Oldfield, R. C. (1971). The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9, 97–113. doi: 10.1016/0028-3932(71)90067-4

PubMed Abstract | Crossref Full Text | Google Scholar

Paczynski, M., and Kuperberg, G. R. (2011). Electrophysiological evidence for use of the animacy hierarchy, but not thematic role assignment, during verb-argument processing. Lang. Cogn. Processes 26, 1402–1456. doi: 10.1080/01690965.2011.580143

PubMed Abstract | Crossref Full Text | Google Scholar

Pesetsky, D. (1995). Zero Syntax—Experiences and Cascades. Cambridge, Mass: MIT Press.

Google Scholar

Pion-Tonachini, L., Kreutz-Delgado, K., and Makeig, S. (2019). ICLabel: An automated electroencephalographic independent component classifier, dataset, and website. NeuroImage 198, 181–197. doi: 10.1016/j.neuroimage.2019.05.026

PubMed Abstract | Crossref Full Text | Google Scholar

Postal, P. M. (1971). Crossover Phenomena. New York: Holt, Reinhart and Winston.

Google Scholar

R Core Team. (2024). R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.

Google Scholar

Rafi, Z., and Greenland, S. (2020). Semantic and cognitive tools to aid statistical science: replace confidence and significance by compatibility and surprise. BMC Med. Res. Methodol. 20, 1–11. doi: 10.1186/s12874-020-01105-9

PubMed Abstract | Crossref Full Text | Google Scholar

Rajendran, S. (1977). Case system in Tamil and Malayalam. Bull. Deccan Coll. Res. Inst. 37, 128–140.

Google Scholar

Rappaport Hovav, M., and Levin, B. (2015). “The syntax-semantics interface,” in The Handbook of Contemporary Semantic Theory (Oxford: Wiley-Blackwell), 593–624. doi: 10.1002/9781118882139.ch19

Crossref Full Text | Google Scholar

Rodríguez-Ferreiro, J., Gennari, S. P., Davies, R., and Cuetos, F. (2011). Neural correlates of abstract verb processing. J. Cogn. Neurosci. 23, 106–118. doi: 10.1162/jocn.2010.21414

PubMed Abstract | Crossref Full Text | Google Scholar

Roehm, D., Winkler, T., Swaab, T., and Klimesch, W. (2002). The N400 and delta oscillations: Is there a difference? J. Cogn. Neurosci. 14, 134–135. doi: 10.1162/089892902317318983

Crossref Full Text | Google Scholar

Schad, D. J., Vasishth, S., Hohenstein, S., and Kliegl, R. (2020). How to capitalize on a priori contrasts in linear (mixed) models: a tutorial. J. Mem. Lang. 110:104038. doi: 10.1016/j.jml.2019.104038

Crossref Full Text | Google Scholar

Schlesewsky, M., and Bornkessel, I. (2006). Context-sensitive neural responses to conflict resolution: electrophysiological evidence from subject–object ambiguities in language comprehension. Brain Res. 1098, 139–152. doi: 10.1016/j.brainres.2006.04.080

PubMed Abstract | Crossref Full Text | Google Scholar

Shalu, S., Muralikrishnan, R., and Choudhary, K. K. (2025). Similar does not mean the same: ERP correlates of mental and physical experiencer verb processing in Malayalam complex constructions. Front. Hum. Neurosci. 19:1632844. doi: 10.3389/fnhum.2025.1632844

PubMed Abstract | Crossref Full Text | Google Scholar

Shannon, C. E. (1948). A mathematical theory of communication. Bell Syst. Tech. J. 27, 379–423. doi: 10.1002/j.1538-7305.1948.tb01338.x

Crossref Full Text | Google Scholar

Shimoyoshi, A. (2015). Psych verbs in Spanish and Japanese: A contrastive study on the semantics-syntax interface (Doctoral dissertation). Kobe City University, Kobe.

Google Scholar

Steinhauer, K., Royle, P., Drury, J. E., and Fromont, L. A. (2017). The priming of priming: evidence that the N400 reflects context-dependent post-retrieval word integration in working memory. Neurosci. Lett. 651, 192–197. doi: 10.1016/j.neulet.2017.05.007

PubMed Abstract | Crossref Full Text | Google Scholar

Subbārāo, K. V. (2012). South Asian Languages. Cambridge: Cambridge University Press.

Google Scholar

Temme, A. (2019). The peculiar nature of psych verbs and experiencer object structures (Doctoral dissertation). Humboldt-Universität Zu, Berlin.

Google Scholar

Tettamanti, M., Buccino, G., Saccuman, M. C., Gallese, V., Danna, M., Scifo, P., et al. (2005). Listening to action-related sentences activates fronto-parietal motor circuits. J. Cogn. Neurosci. 17, 273–281. doi: 10.1162/0898929053124965

PubMed Abstract | Crossref Full Text | Google Scholar

Thompson, C. K., and Lee, M. (2009). Psych verb production and comprehension in agrammatic Broca's aphasia. J. Neurolinguist. 22, 354–369. doi: 10.1016/j.jneuroling.2008.11.003

PubMed Abstract | Crossref Full Text | Google Scholar

Van Valin Jr R. D. (2005). A summary of Role and Reference Grammar (Unpublished manuscript). Available online at: http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/RRGsummary.pdf (Accessed November 28, 2025)

Google Scholar

Van Valin, R. D. Jr. (1991). Another look at Icelandic case marking and grammatical relations. Nat. Lang. Linguist. Theor. 9, 145–194.

Google Scholar

Wang, L., Schlesewsky, M., Bickel, B., and Bornkessel-Schlesewsky, I. (2009). Exploring the nature of the “subject'-preference: evidence from the online comprehension of simple sentences in Mandarin Chinese. Lang. Cogn. Processes 24, 1180–1226. doi: 10.1080/01690960802159937

Crossref Full Text | Google Scholar

Wang, L., and Schumacher, P. B. (2013). New is not always costly: evidence from online processing of topic and contrast in Japanese. Front. Psychol. 4:363. doi: 10.3389/fpsyg.2013.00363

PubMed Abstract | Crossref Full Text | Google Scholar

Wasserstein, R. L., Schirm, A. L., and Lazar, N. A. (2019). Moving to a World Beyond “p < 0.05.” Am. Stat. 73, 1–19. doi: 10.1080/00031305.2019.1583913

Crossref Full Text | Google Scholar

Wilson, M., and Dillon, B. (2022). Alignment between thematic roles and grammatical functions facilitates sentence processing: evidence from experiencer verbs. SSRN Electron. J. doi: 10.2139/ssrn.4235952

Crossref Full Text | Google Scholar

Wunderlich, D. (1997). Cause and the structure of verbs. Ling. Inquiry 28, 27–68.

Google Scholar

Wunderlich, D. (2003). “Optimal case patterns: German and Icelandic compared,” in New Perspectives on Case Theory, eds. E. Brandner and H. Zinsmeister (Stanford: CSLI Publications), 329–365.

Google Scholar