648844 research-article2016

FLA0010.1177/0142723716648844First LanguageLalioti et al.

FIRST LANGUAGE

Article

Subject–verb agreement and verbal short-term memory: A perspective from Greek children with specific language impairment

First Language 2016, Vol. 36(3) 279­–294 © The Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0142723716648844 fla.sagepub.com

Marina Lalioti, Stavroula Stavrakaki, Christina Manouilidou and Ioanna Talli

Aristotle University of Thessaloniki, Greece; University of Patras, Greece

Abstract This study investigated the performance of school age Greek-speaking children with SLI on verbal short-term memory (VSTM) and Subject–Verb (S-V) agreement in comparison to chronological age controls and younger typically developing children. VSTM abilities were assessed by means of a non-word repetition task (NRT) and an elicited production task, an off-line grammaticality judgment (GJ) task, and an on-line self-paced listening task were employed to examine S-V agreement abilities. The study found significant between-group differences in the NRT and the off-line GJ task. No significant between-group differences were found for the S-V agreement production task and for reaction times of the on-line task. The authors argue that the children with SLI showed increased grammatical sensitivity in S-V agreement and relate their lower performance on the off-line GJ task to the task’s metalinguistic demands. The findings suggest that deficits in VSTM constitute a robust marker for SLI and claim that S-V agreement can be acquired by SLI individuals with apparent limitations in VSTM. Keywords Grammatical sensitivity, limitations in verbal short-term memory, specific language impairment, subject–verb agreement, verbal short-term memory

Corresponding author: Stavroula Stavrakaki, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece. Email: [email protected]

Downloaded from fla.sagepub.com by guest on June 10, 2016

280

First Language 36(3)

Introduction Recent studies on Specific Language Impairment (SLI), a condition that primarily affects children’s linguistic abilities without apparent intellectual limitations, or physical or sensory deficits (Leonard, 2014), have revealed deficits in verb agreement morphology and verbal short-term memory (Archibald & Gathercole, 2007; Botting & Conti-Ramsden, 2001; Botting, Psarou, Caplin, & Nevin, 2013; Clahsen, Bartke, & Göllner, 1997; ContiRamsden, Botting, & Faragher, 2001; Conti-Ramsden & Durkin, 2007; Girbau & Schwartz, 2007; Rice & Bode, 1993; Rice & Oetting, 1993; Rothweiler, Chilla, & Clahsen, 2011; van der Lely & Marshall, 2011; van der Lely & Ullman, 1996). A large number of studies on different languages have investigated the performance of children with SLI on subject–verb agreement and on verbal short-term memory tasks.

Verb agreement morphology in children with SLI Cross-linguistic findings show limitations in the domain of subject–verb agreement (S-V agreement). The bulk of research concerns studies on English, a language with sparse inflections. The findings show that English-speaking children with SLI show severe difficulties in the acquisition of the verb agreement morpheme –s which can persist into young adulthood (Clahsen et al., 1997; Leonard, 2009, 2014; van der Lely, 1998). Cross-linguistic evidence on German and Dutch corroborate the findings that the acquisition of S-V agreement is particularly challenging for children with SLI (Clahsen et al., 1997; Leonard, 2009, 2014; Rothweiler et al., 2011; Verhagen & Blom, 2014). However, as pointed out by Leonard (2009, pp. 312–313), the differences between German and Dutch children with SLI on verb inflectional morphemes with their language-matched peers ‘do not appear to be of the same magnitude as is found in English’ as the typology of verb second languages such as German and Dutch is different and ‘inflections are greater in number in these languages than in English’. Additional evidence confirming the specific language typology effect comes from null subject languages with rich inflection, such as Italian and Spanish. Findings from these languages indicate that children with SLI do take ‘the rich inflection advantage’ and show fewer differences from their peers on the production and comprehension of verb agreement morphemes (Leonard, 2009, 2014).

Verbal short-term memory in SLI A large number of studies (e.g. Girbau & Schwartz, 2007, 2008; Poll, Betz, & Miller, 2010; Rispens & Been, 2007; Smith, 2008; Talli, 2010; Thordardottir & Brandeker, 2013; Vandewalle, Boets, Ghesquière, & Zink, 2010) have employed non-word repetition tasks (NRT) to measure verbal short-term memory (VSTM), as performance on NRT relies heavily on the use of the phonological short-term store, i.e. the mechanism used for the phonological representation of information in working memory (Gathercole & Baddeley, 1990; Gathercole, Willis, Baddeley, & Emslie, 1994; Girbau & Schwartz, 2008; Thordardottir & Brandeker, 2013). These studies indicated that children with SLI

Downloaded from fla.sagepub.com by guest on June 10, 2016

281

Lalioti et al.

usually have limited VSTM capacity that in turn leads to major difficulties with the recall and the repetition of non-words. While most of studies investigate the performance of children with SLI on VSTM and S-V agreement (or more generally, verb inflection) separately, a few have provided simultaneous examination of these markers of SLI. Among the first researchers to do so were Conti-Ramsden et al. (2001), who assessed the abilities of children with SLI in an NRT, a 3rd person singular task, a past tense task, and a sentence repetition task. Sentence repetition was the most sensitive psycholinguistic marker for SLI and as such the most useful marker. Rispens and Been (2007) found that typically developing children outperform children with SLI on tasks of S-V agreement and non-word repetition. The investigators argue for the co-occurrence of these difficulties (Rispens & Been, 2007, p. 294). Bishop, Adams, and Norbury (2006), in contrast, found that children with SLI demonstrate deficits in VSTM and/or S-V agreement. That is, in SLI, VSTM problems are sometimes, but not always, accompanied by S-V agreement deficits, and vice versa (Bishop et al., 2006).

S-V agreement and VSTM in Greek children with SLI S-V agreement in SLI has been at the center of research with Greek-speaking children. The majority of studies have focused on preschool children and revealed consistent deficits in S-V agreement especially in marking the 2nd person singular and the 2nd person plural. (An overview of the studies on S-V agreement in Greek SLI are presented in Appendix I in the online supplement.) The studies with Greek children with SLI are small in scale: case studies and investigations with a small number of participants (Clahsen & Dalalakis, 1999; Mavratza, 2007; Stavrakaki, 1996; Tsimpli, 2001; Varlokosta, 2000a, 2000b; for reviews see, Stavrakaki, 2005; Stavrakaki, Chrysomallis, & Petraki, 2011). Interestingly, even with this small body of research, some findings have been consistently observed. First, young children with SLI perform poorly on the 2nd person singular and plural (Stavrakaki, Vogindroukas, Chelas, & Ghousi, 2008). Second, they perform well (near ceiling or ceiling) on the 3rd person singular (suffix –i). The overuse of the verbal suffix –i (identical to the 3rd person singular ending) has been repeatedly observed by Greek L1 acquisition researchers (Katis, 1984; Stavrakaki & Okalidou, 2016; Stephany, 1997;Varlokosta, 2005; Varlokosta, Vainikka, & Rohrbacher, 1998). Smith (2008) employed an off-line elicited production task and showed that children with SLI aged 4;6–6;9 performed like their language age (LA) controls on the production of S-V agreement but importantly significantly below their chronological age (CA) controls. Children’s performances on 2nd person singular, 2nd person plural, and 1st person plural were remarkably low (incorrect performance: 42%, 20%, and 29% respectively). The performance of preschool children with SLI has been found to be consistently poor compared to that of typically developing Greek children who demonstrate knowledge of the complex Greek verb morphology by around 3 years of age (Katis, 1984; Stavrakaki & Okalidou, 2016; Stephany, 1997). Nonetheless, inconsistencies have also been noted. Data from older children with SLI show that deficits in the domain of S-V agreement decrease with age. Stavrakaki (2001) observed that school age children with SLI performed well in this domain while at the

Downloaded from fla.sagepub.com by guest on June 10, 2016

282

First Language 36(3)

same time demonstrating severe impairment in elicited production and comprehension tasks of complex syntactic structures. Stavrakaki et al. (2011) investigated the performance of a French–Greek simultaneous bilingual child with SLI (aged 9 years) in various linguistic tasks. These researchers found that the child performed almost at ceiling, and exactly like two comparison typically developing (matched on language age) bilingual participants, on an S-V elicitation task for Greek but not for French. This case study indicates that the S-V agreement acquisition is more easily achieved by children with SLI in a language with rich inflection (Greek) than in a weakly inflected language (for example, French; Dressler, 2007). Few studies, however, have investigated VSTM in Greek children with SLI. Stavrakaki (2001) found that school age children with SLI performed below their CA controls on an NRT. Similarly, Smith (2008) found that children with SLI aged 4;9–6;9 had significant difficulties in an NRT and a digit span task, compared to their LA and CA matched peers. Talli (2010) measured and compared the performance of 8- to 10-year-old Greekspeaking children in an NRT. She found that children with SLI were significantly outperformed by their peers. However, they differed only in terms of accuracy and not speed (Talli, 2010). Smith (2008) examined S-V agreement and VSTM simultaneously. She found that performance on VSTM is a better clinical indicator of Greek SLI than S-V agreement, at least in preschool children. In contrast, Talli (2010) and Talli, Sprenger-Charolles, and Stavrakaki (2013, 2016) found that Greek school children with SLI demonstrated more severe deficits in morphosyntax as measured by the Greek version of the Test for Reception of Grammar (TROG), than in an NRT. These inconsistent findings indicate that the relation between VSTM and morphosyntax warrants further examination. In this investigation, we examine S-V agreement abilities comprehensively by employing tasks not previously employed in this area of research alongside VSTM as measured by NRT in a sample of Greek school age children with SLI. The study aims to shed light on the relationship between these abilities in school age children with SLI in the context of Greek, a highly inflected language.

Method Participants In total, 46 children participated in the study. There were 10 Greek native speaking children with SLI (mean age 8;5 years). At the time of testing two children were exposed to another language in addition to Greek at home (one to Romanian and another one to Russian). In addition, all children above the age of 8 years were exposed to English at school. The children were diagnosed with SLI by clinicians and multidisciplinary groups in Thessaloniki on the basis of the following criteria: normal non-verbal IQ assessed by qualified psychologists (see Appendix II, online supplementary materials for the non-verbal IQ scores); absence of sensory, motor or other disorders (e.g. ADHD, autism); absence of severe environmental deprivation (language, socialization, etc.) as reported by teachers or parents. All these children showed apparent limitation in language abilities that were further assessed by the means of the Greek version of the

Downloaded from fla.sagepub.com by guest on June 10, 2016

283

Lalioti et al.

Table 1.  Number of participants, chronological age (Mean and SD) and performance of each group (Mean % and SD) on the Diagnostic Verbal IQ test and The Renfrew Action Picture Test (Greek version). Group

SLI TD_YC TD_CA

N

10 24 12

CA

DVIQ for preschool children

8;5 (1;5) 4;11 (0;3) 9;7 (1;6)

RAPT

Vocabulary

Production of morphosyntax

Sentence repetition

61.1 (8.4) 47.1 (10.0) –

64.3 (21.8) 46.1 (11.3) –

86.5 (10.1) 93.3 (5.6) –

78.5 (6.5) 74.1 (8.6) –

SLI: specific language impairment group; CA: chronological age; TD_YC: typically developing younger controls; TD_CA: typically developing chronological age controls.

Renfrew Action Picture Test (RAPTGR, Vogindroukas, Protopapas, & Stavrakaki, 2009) as well as the production of vocabulary, morphosyntax, and sentence repetition subsections of the Greek Diagnostic Verbal IQ Test for preschool children (DVIQ, Stavrakaki & Tsimpli, 2000). Comparison groups comprised 24 younger typically developing children (TD_YC; mean age 4;11 years) and 12 typically developing children matched for chronological age to the children with SLI (TD_CA; mean age 9;7 years). Table 1 presents the characteristics of the three groups of children participating in the study. There was no significant difference between the children with SLI and younger controls on the sentence repetition (normal distribution was shown by Kolmogorov–Smirnoff test, p > .05; t (32) = 2.030, p = .066) as well as on RAPTGR (Mann–Whitney U, Z = .834, p = .423). However, the participants with SLI performed better on the vocabulary and morphosyntax part of the preschool version of the DVIQ test (Mann–Whitney U, Z = 3.448, p = .001; and Z = 2.169, p = .03 respectively). This might be expected as this version of the DVIQ was developed for preschoolers while the SLI participants were at school age and receiving therapy services at the time of testing.

Materials and procedure In total, four tasks were employed. Specifically, the children were tested on an NRT and three tasks tapping S-V agreement: (1) an off-line elicited production task, (2) an off-line grammatical judgment (GJ) task, and (3) an on-line self-paced listening task. Each of the three S-V agreement tasks tested a different aspect of the children’s linguistic abilities in this domain. In particular, the off-line GJ task tested metalinguistic abilities while the on-line S-V agreement task tested the implicit knowledge of this grammatical phenomenon. The elicited production task tested the children’s oral expressive abilities regarding S-V agreement. All children were individually tested by the first author of this study (ML), either at their homes or at the practitioners’ offices and schools. The researcher met each child one to three times (total duration of meetings: 30–90 minutes with breaks). Notably, the younger controls were not able to engage with two tasks: (1) the grammaticality judgment task, which requires advanced metalinguistic abilities acquired late in childhood

Downloaded from fla.sagepub.com by guest on June 10, 2016

284

First Language 36(3)

Table 2.  Number of participants in each task and their mean chronological age (SD). Group

DVIQ/RAPT

Non-word repetition task

Off-line elicited production task

Off-line grammaticality judgment task

On-line self-paced listening task

SLI

10/ 8;5 (1;5) 24/ 4;11 (0;3) 0/ 0

7/ 8;10 (1;6) 24/ 4;11 (0;3) 12/ 9;7 (1;6)

7/ 8;10 (1;6) 24/ 4;11 (0;3) 12/ 9;7 (1;6)

10/ 8;5 (1;5) 0/ 0

6/ 9;1 (1;6) 0/ 0

11/ 9;10 (1;4)

9/ 9;8 (1;5)

TD_YC TD_CA

DVIQ: Diagnostic Verbal IQ Test; RAPT: Renfrew Action Picture Test; SLI: specific language impairment group; TD_YC: typically developing younger controls; TD_CA: typically developing chronological age controls.

(McDaniel & Cairns, 1996); and (2) the on-line task, which the younger TD children were also unable to follow. Some children with SLI were unable to perform on some of the tasks. These children were excluded from the analyses. Table 2 presents the final number and age profile of the participants in each group and task.

Non-word repetition task We employed the NRT developed by Talli (2010). Children listened and then repeated one by one a total of 24 non-words orally presented by the experimenter, varying in length from three to six syllables with a CV and CVC syllable structure, e.g. mukola, vardotivaruto. There were six non-words for each length of which three included only CV syllables, while the other three included a CVC syllable presented in increasing order of length. For this task, there are normative data from 174 Greek-speaking children: 54 children 7–8 years old; 64 children 8–9 years old; 56 children 9–10 years old. The expected mean percentage correct responses and SDs for typical developing children in each age range are available in Talli (2010).

Off-line elicited production task The off-line elicited production task (Stavrakaki et al., 2011) was employed in order to test the children’s S-V agreement production abilities. The task was orally presented by the experimenter. In this task, children were asked to conjugate 10 verbs in an indirect manner, with the use of some props and imaginary settings. For instance, when the aim of the task was to elicit the 2nd person singular of the verb ‘to read’, the experimenter would hold a book and pretend that she was reading. Once this was done, the experimenter would ask the child ‘What am I doing now?’ In total, 60 verb forms were elicited per child.

Downloaded from fla.sagepub.com by guest on June 10, 2016

285

Lalioti et al.

Off-line grammatical judgment task The purpose of the off-line GJ task was to check the children’s S-V agreement judgment abilities. In the off-line GJ task, the children had to listen to and judge the correctness of 52 sentences (36 experimental and 16 filler sentences). All the experimental sentences had the same basic structure, Subject+Verb+Object+Adverbial/Prepositional Phrase. Half of them were grammatically correct, whereas the remaining ones contained a S-V agreement violation (see example in 1(a) below). As regards the filler sentences, half of them semantically correct and the other half contained a semantic violation (see Appendix III, online supplement). The purpose of the latter sentences was to ensure that the participants paid attention to the task and did not give random answers. It should also be noted that many different variables were controlled (such as the length of the sentences in terms of syllable number, the distance between the subject and the verb, etc.), in order to minimize confounding effects on the children’s performance. The experimental sentences were divided into two different lists, so that almost half of the children encountered the grammatical version of a specific experimental sentence, while the other half encountered its ungrammatical counterpart. The sentences were presented orally by the experimenter. Experimental sentence 1

(a)Aftos-3SG *vafun-3PL to ðomatio me boʝa He *paint the room with paint

On-line self-paced listening task The purpose of the task was to examine children’s S-V agreement judgment abilities, but in this case, without the assistance of their metalinguistic knowledge and the confounding factor of working memory. The same experimental sentences as those used in the off-line GJ task were used (see example in 2(a) below and Appendix III, supplementary materials) whereas the filler sentences were 18 pairs of affirmative and interrogative sentences. The on-line task was presented on a laptop computer, using internal speakers and keyboard. The children were asked to listen to 72 sentences (36 experimental and 36 filler sentences), presented to them segment-by-segment from a laptop computer. Auditory materials were recorded and then adjusted by using Praat (Boersma, 2001), while for stimulus presentation and reaction times (RTs) recording E-Prime 2.0. professional was used (Schneider, Eschman, & Zuccolotto, 2002a, 2002b). Finally, similarly to the off-line GJ, the experimental sentences were divided into two different lists, so that almost half of the children encountered the grammatical version of a specific experimental sentence, while the other half encountered its ungrammatical counterpart. The task was completed after the presentation of all sentences. Children were instructed to listen to each segment and press the space bar to move on to the following segment and continue listening. In order to assure children’s engagement in the task, YES or NO comprehension questions appeared after the filler sentences. An example of the presentation procedure is presented in 2(b) below. Materials were presented in random order.

Downloaded from fla.sagepub.com by guest on June 10, 2016

286

First Language 36(3)

Experimental sentence and presentation procedure 2 (a) Aftos-3SG *vafun-3PL to ðomatio me boʝa *paint the room with paint He (b) Aftos-3SG (the participant had to press SPACE in order to hear the next segment) (the participant had to press SPACE in order to hear the next segment) *vafun-3PL to ðomatio (the participant had to press SPACE in order to hear the next segment) me boʝa (the participant had to press SPACE in order to hear the next segment)

Results Non-word repetition task The performance of all groups on the NRT is shown in Table 3. As expected, the children with SLI scored lower than both comparison groups of typically developing peers. A one-way analysis of variance indicated a significant group effect on the NRT (F (2, 40) = 6.833, p = .003). Post-hoc comparisons revealed the children with SLI performed significantly below their CA controls (Z = 2.934, p = .003) while their scores were not significantly different from those of their younger TD peers (Z = 1.473, p = .153).

Off-line elicited production task The performance of all groups on the off-line S-V agreement production task is shown in Table 3. Groups exhibited high performance, with CA controls performing at ceiling. Similar overall performance was also evident for children with SLI and their younger peers. A one way ANOVA did not show a significant group effect on the production of verb agreement morphemes (F (2, 40) = 1.679, p = .199).

Off-line grammaticality judgment task Within-group analysis indicated that the overall performance of the SLI children on the condition with grammatical sentences (grammatical condition) did not differ from the Table 3.  Performance (mean % and SD) of each group on the non-word repetition task and the S-V agreement off-line elicited production task. Group NRT S-V agreement Total 1st person 2nd person 3rd person 1st person 2nd person 3rd person score singular singular singular plural plural plural SLI TD_ YC TD_ CA

62 (16) 72 (16) 86 (10)

99 (2) 98 (3) 100 –

100 – 100 – 100 –

97 (8) 100 – 100 –

100 – 99 (2) 100 –

100 – 98 (5) 100 –

96 (5) 93 (17) 100 –

100 – 99 (2) 100 –

NRT: non-word repetition task; S-V: subject-verb; SLI: specific language impairment group; TD_YC: typically developing younger controls; TD_CA: typically developing chronological age controls.

Downloaded from fla.sagepub.com by guest on June 10, 2016

287

Lalioti et al.

Table 4.  Performance (mean % and SD) of each group on the off-line grammaticality judgment task: grammatical and ungrammatical condition. Group

Total score 1st person 2nd person 3rd person 1st person 2nd person 3rd person singular singular singular plural plural plural

Grammatical condition SLI 78 (15) 93 (14) TD_CA 99 (2) 100 Ungrammatical condition SLI 56 (32) 53 (45) TD_CA 99 (2) 97 (10)

77 (23) 97 (10)

67 (35) 100

83 (18) 100

73 (26) 97 (10)

77 (27) 100

50 (45) 100

53 (36) 100

67 (35) 100

67 (31) 97 (10)

46 (45) 100

SLI: specific language impairment group; TD_CA: typically developing chronological age controls.

condition with the ungrammatical ones (ungrammatical condition), (t (9) = 2.46, p = .071), and thus there was no grammaticality effect on the performance of children with SLI. The CA controls performed at ceiling on both the grammatical and ungrammatical condition indicating that the children had acquired grammatical abilities tapped by this task. As noted earlier, the younger TD children were unable to engage with the off-line elicited production task. Details of comparisons between the children with SLI and TD_ CA on the grammatical and ungrammatical condition are presented in Table 4. As seen in Table 4, TD_CA children outperformed the children with SLI on the grammatical judgment of the grammatically correct sentences (total score and all verb person, t tests, p < .05, except for the 1st person singular, t (19) = 1.5, p = .168). One only child with SLI performed at ceiling. In contrast, all the TD_CA children completed the task without errors (100%) with only two children performing at 94%.

On-line self-paced listening task Reaction times for all four sentence segments (Subject, Verb, Object, and Adverbial/ Prepositional Phrase) constituted the data for analyses. For the item analyses, the data were the means of each of the four segments of the experimental sentences. Prior to the analyses, RTs above 1800 ms were removed and replaced with the mean RT of each segment. Also, outliers equal to two standard deviations below or above the mean for each segment were also removed and replaced with the mean. The younger TD children could not follow the task. We thus report results from two groups (TD_CA children and SLI children). The results for the overall group performance on all verb persons in the grammatical and ungrammatical condition are presented in Figure 1 (grammatical) and Figure 2 (ungrammatical). No between-group differences in RTs were found although children in both groups had higher RTs in the ungrammatical condition compared to the grammatical condition. A 2 (group: SLI vs. TD_CA) × 2 (condition: grammatical vs. ungrammatical) analysis of variance (ANOVA) revealed no effect of group (F (1, 13) = .008, p = .930) while condition was significant (F (1, 13) = 43.822, p < .001) reflecting the faster RTs by both groups in the grammatical than the ungrammatical condition. The interaction between group and

Downloaded from fla.sagepub.com by guest on June 10, 2016

288

First Language 36(3)

Figure 1.  Mean reaction times per group, per segment in the grammatical condition.

Figure 2.  Mean reaction times per group, per segment in the ungrammatical condition.

condition was also not significant, indicating that both groups showed the same performance pattern (F (1, 13) = .052, p = .824). Table 5 presents the findings in detail.

Discussion This study was designed to simultaneously investigate the performance of school age Greek-speaking children with SLI on verbal short-term memory (VSTM) and subject– verb (S-V) agreement. We employed a non-word repetition task to assess the VSTM abilities of these children and three different S-V agreement tasks which examined S-V

Downloaded from fla.sagepub.com by guest on June 10, 2016

289

Lalioti et al. Table 5.  Reaction times in milliseconds (and SD) of each group on the on-line self-paced listening task: verb segment. Group

Total score

Grammatical condition SLI 1052 (233) TD_CA 1042 (245) Ungrammatical condition SLI 1103 (244) TD_CA 1090 (261)

1st person singular

2nd person singular

3rd person singular

1st person plural

2nd person plural

3rd person plural

963 (175) 1031 (203)

1007 (234) 986 (227)

1015 (314) 1105 (243) 961 (228) 1100 (266)

1094 (268) 1102 (353)

1128 (276) 1073 (283)

1038 (286) 1038 (241)

1127 (306) 1055 (289)

1117 (208) 1078 (221) 1134 (281) 1062 (268)

1110 (285) 1126 (289)

1149 (226) 1123 (250)

SLI: specific language impairment group; TD_CA: typically developing chronological age controls.

agreement from different angles: (1) an off-line elicited production task, (2) an off-line grammaticality judgment (GJ) task in which half of the experimental sentences contained a S-V agreement violation, and (3) an on-line self-paced listening task in which half of the test sentences contained a S-V agreement violation.

Subject-verb agreement and verbal short-term memory in Greek children with SLI The picture regarding S-V agreement by children with SLI is complex. On the one hand, no significant group effect was found on the oral production of S-V agreement. In addition, the children with SLI performed as their chronological age (CA) peers on the online processing of S-V agreement. Reaction times were longer for ungrammatical than grammatical sentences in both groups. On the other hand, performance on the GJ task was significantly below that of typically developing peers. The findings reveal that school age Greek children with SLI are capable of acquiring S-V agreement; they perform at the same level as their CA peers in this domain. Our own previous research in addition to findings from other studies involving Greek children with SLI indicate significant difficulties for preschoolers with SLI and performance below their CA peers especially in marking the 2nd singular and plural person in their spontaneous speech (see, Stavrakaki, 2005, for a review of the literature; Mavratza, 2007; Smith, 2008; Stavrakaki et al., 2008). The present study suggests that for older, school age children with SLI, S-V agreement is no longer a major challenge. We found good performance by children with SLI even on the verb persons which are difficult to acquire (2nd person singular and plural). This finding contrasts with findings especially from English children with SLI which indicate that deficits in the verbal domain (including those of agreement) persist into young adulthood (van der Lely, 1998, 2005). We can thus conclude that the linguistic input and the ‘surface’ properties of each language can significantly affect the linguistic development of children with SLI (Leonard, 2009, 2014). It seems that the rich inflectional properties of Greek facilitate the acquisition of a local dependency, namely S-V agreement. The performance of children with SLI on the GJ task is not parallel to their production and on-line processing abilities. Apparently, the off-line GJ task presents significant

Downloaded from fla.sagepub.com by guest on June 10, 2016

290

First Language 36(3)

difficulties to children with SLI. With respect to typical development, it has been argued by Cairns, Schlisselberg, Waltzman, and McDaniel (2006) that the ability to make grammaticality judgments requires the child to access and employ syntactic knowledge explicitly. This ability is acquired late by typically developing children (Cairns et al., 2006). With respect to children with SLI, a number of cross-linguistic findings indicate difficulties in performing GJ tasks (Jung & Pae, 2010; Maillart, Schelstraete, & Hupet, 2004; Rice, Hoffman, & Wexler, 2009; Rice, Wexler, & Redmond, 1999; van der Lely & Marshall, 2011; van der Lely & Ullman, 1996). Our findings, in addition to giving support to previous studies showing difficulties in the judgment abilities of the children with SLI, reveal a between-task dissociation (elicited production and on-line task vs. off-line GJ task). This dissociation may be due to the fact that S-V agreement per se is not difficult to acquire in Greek, as discussed above, while explicit or conscious access to grammatical knowledge develops independently to some extent. In terms of VSTM, the performance of children with SLI was significantly below that of their CA peers but at similar levels to that of their younger typically developing peers. These findings indicate a significant developmental delay for the individuals with SLI. In this respect, they confirm that impairments in VSTM constitute a robust marker for Greek SLI (Archibald & Gathercole, 2007; Girbau & Schwartz, 2007, 2008; Poll et al., 2010; Rispens & Been, 2007; Smith, 2008).

The relation between VSTM and S-V agreement in Greek SLI The results indicated a notable dissociation between VSTM abilities on the one hand and S-V agreement marking abilities on other hand, as assessed by the elicited production task and the timed on-line task. This dissociation warrants further investigation. A number of researchers (Botting & Conti-Ramsden, 2001; Rispens & Been, 2007) report parallel deficits in these domains in Dutch and English children with SLI and postulate that these domains are closely related and develop interdependently. In our view, our findings suggest that a language with rich inflection can facilitate the acquisition of S-V agreement. In particular, our results call into question the close relationship between these abilities and indicate that in certain linguistic environments they can develop independently. Further evidence for this interpretation comes from the work of Bishop et al. (2006), who found minimal correlation between phonological STM deficits and morphological (tense) marking deficits in a large SLI sample. In the same vein, Smith (2008) showed that performance on S-V agreement cannot be predicted by performance on the NRT in a preschool sample of children with SLI. Like Noonan, Redmond, and Archibald (2014) we suggest that working memory and linguistic skills may not impact in the same way on the metalinguistic judgments of children with language impairment and children with both language impairment and memory disability. Collectively, such findings suggest that the relationship between S-V agreement and VSTM is not as close as previously suggested. However, it needs to be acknowledged that the evidence from this investigation is based on a small sample of children and further research is required to replicate these findings with larger samples. Future research would also benefit from directly comparing children with SLI exposed to languages with different inflectional morphosyntactic characteristics.

Downloaded from fla.sagepub.com by guest on June 10, 2016

291

Lalioti et al. Acknowledgements

We would like to thank Ana Vivas for her technical assistance. We are also very grateful to Gina Conti-Ramsden, Chloe Marshall, and the two anonymous reviewers for comments that greatly improved our manuscript.

Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

References Archibald, L. M., & Gathercole, S. E. (2007). Nonword repetition in specific language impairment: More than a phonological short-term memory deficit. Psychonomic Bulletin & Review, 14, 919–924. Bishop, D. V. M., Adams, C. V., & Norbury, C. F. (2006). Distinct genetic influences on grammar and phonological short-term memory deficits: Evidence from 6-year-old twins. Genes, Brain and Behavior, 5, 158–169. Boersma, P. (2001). Praat, a system for doing phonetics by computer. Glot International, 5, 341–345 Botting, N., & Conti-Ramsden, G. (2001). Non-word repetition and language development in children with specific language impairment (SLI). International Journal of Language & Communication Disorders, 36, 421–432. Botting, N., Psarou, P., Caplin, T., & Nevin, L. (2013). Short-term memory skills in children with specific language impairment: The effect of verbal and nonverbal task content. Topics in Language Disorders, 33, 313–327. Cairns, H. S., Schlisselberg, G., Waltzman, D., & McDaniel, D. (2006). Development of a metalinguistic skill judging the grammaticality of sentences. Communication Disorders Quarterly, 27, 213–220. Clahsen, H., Bartke, S., & Göllner, S. (1997). Formal features in impaired grammars: A comparison of English and German SLI children. Journal of Neurolinguistics, 10, 151–171. Clahsen, H., & Dalalakis, J. (1999). Tense and agreement in Greek SLI: A case study. Colchester, UK: Department of Language and Linguistics, University of Essex. Conti-Ramsden, G., Botting, N., & Faragher, B. (2001). Psycholinguistic markers for specific language impairment (SLI). Journal of Child Psychology and Psychiatry, 42, 741–748. Conti-Ramsden, G., & Durkin, K. (2007). Phonological short-term memory, language and literacy: Developmental relationships in early adolescence in young people with SLI. Journal of Child Psychology and Psychiatry, 48, 147–156. Dressler, W. U. (2007). Introduction. In S. Laaha & S. Gillis (Eds.), Typological perspectives on the acquisition of noun and verb morphology (Antwerp Papers in Linguistics, Vol. 112, pp. 3–9). Antwerp, Belgium: University of Antwerp. Gathercole, S. E., & Baddeley, A. D. (1990). Phonological memory deficits in language disordered children: Is there a causal connection? Journal of Memory and Language, 29, 336–360. Gathercole, S. E., Willis, C. S., Baddeley, A. D., & Emslie, H. (1994). The children’s test of nonword repetition: A test of phonological working memory. Memory, 2, 103–127.

Downloaded from fla.sagepub.com by guest on June 10, 2016

292

First Language 36(3)

Girbau, D., & Schwartz, R. G. (2007). Non-word repetition in Spanish-speaking children with specific language impairment (SLI). International Journal of Language & Communication Disorders, 42, 59–75. Girbau, D., & Schwartz, R. G. (2008). Phonological working memory in Spanish–English bilingual children with and without specific language impairment. Journal of Communication Disorders, 41, 124–145. Jung, K., & Pae, S. (2010). The grammaticality judgment in school-aged children with specific language impairment. Communication Sciences & Disorders, 15, 619–631. Katis, D. (1984). The acquisition of the modern Greek verb: With special reference to the imperfective past and perfect classes (Unpublished doctoral dissertation). University of Reading, Reading, UK. Leonard, L. B. (2009). Cross-linguistic studies of child language disorders. In R. G. Schwartz (Ed.), Handbook of child language disorders (pp. 308–324). New York, NY: Psychology Press. Leonard, L. B. (2014). Children with specific language impairment. Cambridge, MA: MIT Press. Maillart, C., Schelstraete, M. A., & Hupet, M. (2004). Phonological representations in children with SLIA study of French. Journal of Speech, Language, and Hearing Research, 47, 187– 198. Mavratza, S. (2007). Specific language impairment in Greek: A study of agreement, tense and the determiner system (Unpublished doctoral dissertation). University of Essex, Colchester, UK. McDaniel, D., & Cairns, H. S. (1996). Eliciting judgments of grammaticality and reference. In D. McDaniel, C. McKee, & H. S. Cairns (Eds.), Methods for assessing children’s syntax (pp. 233–254). Cambridge, MA: MIT Press. Noonan, N. B., Redmond, S. M., & Archibald, L. M. (2014). Contributions of children’s linguistic and working memory proficiencies to their judgments of grammaticality. Journal of Speech, Language, and Hearing Research, 57, 979–989. Poll, G. H., Betz, S. K., & Miller, C. A. (2010). Identification of clinical markers of specific language impairment in adults. Journal of Speech, Language, and Hearing Research, 53, 414–429. Rice, M. L., & Bode, J. (1993). GAPS in the verb lexicons of children with specific language impairment. First Language, 13, 113–131. Rice, M. L., Hoffman, L., & Wexler, K. (2009). Judgments of omitted BE and DO in questions as extended finiteness clinical markers of specific language impairment (SLI) to 15 years: A study of growth and asymptote. Journal of Speech, Language, and Hearing Research, 52, 1417–1433. Rice, M. L., & Oetting, J. B. (1993). Morphological deficits of children with SLI evaluation of number marking and agreement. Journal of Speech, Language, and Hearing Research, 36, 1249–1257. Rice, M. L., Wexler, K., & Redmond, S. M. (1999). Grammaticality judgments of an extended optional infinitive grammar: Evidence from English-speaking children with specific language impairment. Journal of Speech, Language, and Hearing Research, 42, 943–961. Rispens, J. E., & Been, P. H. (2007). Subject-verb agreement and phonological processing in developmental dyslexia and SLI: A closer look. International Journal of Language & Communication Disorders, 42, 293–305. Rothweiler, M., Chilla, S., & Clahsen, H. (2011). Subject-verb agreement in specific language impairment: A study of monolingual and bilingual German-speaking children. Bilingualism: Language and Cognition, 15, 39–57. Schneider, W., Eschman, A., & Zuccolotto, A. (2002a). E-prime user’s guide. Pittsburgh, PA: Psychology Software Tools.

Downloaded from fla.sagepub.com by guest on June 10, 2016

293

Lalioti et al.

Schneider, W., Eschman, A., & Zuccolotto, A. (2002b). E-prime reference guide. Pittsburgh, PA: Psychology Software Tools. Smith, N. (2008). Morphosyntactic skills and phonological short-term memory in Greek preschool children with specific language impairment (Unpublished doctoral dissertation). University of Reading, Reading, UK. Stavrakaki, S. (1996). Specific language impairment in Greek: Evaluation of person and number agreement, case assignment to overt subject pronouns and tense marking (Unpublished master’s thesis). University of Essex, Colchester, UK. Stavrakaki, S. (2001). Specific language impairment in Greek: Aspects of syntactic production and comprehension (Unpublished doctoral dissertation). Aristotle University of Thessaloniki, Thessaloniki, Greece. Stavrakaki, S. (2005). Greek neurolinguistics: The state of the art. Journal of Greek Linguistics, 6, 187–234. Stavrakaki, S., Chrysomallis, M.-A., & Petraki, E. (2011). Subject-verb agreement, object clitics and Wh-questions in bilingual French-Greek SLI: The case study of a French-Greek-speaking child with SLI. Clinical Linguistics and Phonetics, 25, 339–367. Stavrakaki, S., & Okalidou, A. (2016). Gr-Larsp: Towards a Greek version of LARSP. In P. Fletcher, M. Ball & D. Crystal (Eds.), Profiling grammar: More languages of LARSP (pp. 27–41). Bristol, UK: Multilingual Matters. Stavrakaki, S., & Tsimpli, I. M. (2000). Diagnostic verbal IQ test for Greek preschool and school age children: Standardization, statistical analysis, psychometric properties. In The Proceedings of the 8th Symposium of the Panhellenic Association of Logopedists (pp. 95– 106). Athens, Greece: Ellinika Grammata. Stavrakaki, S., Vogindroukas, I., Chelas, E., & Ghousi, S. (2008). Subject-verb agreement in SLI and child L2: A comparative approach to Greek SLI and (un) impaired child L2. In A. Gavarró & M. João Freitas (Eds.), Language acquisition and development. Cambridge, UK: Cambridge Scholars Press (CSP). Stephany, U. (1997). The acquisition of Greek. In D. Slobin (Ed.), The cross linguistic study of language acquisition (Vol. 4, pp. 183–303). Hillsdale, NJ: Lawrence Erlbaum. Talli, I. (2010). Linguistic abilities in developmental dyslexia and specific language impairment (SLI): A comparative and cross-linguistic approach (Unpublished doctoral dissertation). Université Paris Descartes, Paris, France. Talli, I., Sprenger-Charolles, L., & Stavrakaki, S. (2013). Phonological and morpho-syntactic abilities in children with developmental dyslexia and specific language impairment: Evidence from Greek. In S. Stavrakaki, P. Konstantinopoulou, & M. Lalioti (Eds.), Advances in language acquisition (pp. 444–453). Cambridge, UK: Cambridge Scholars Press (CSP). Talli, I., Sprenger-Charolles, L., & Stavrakaki, S. (2016). Specific language impairment and developmental dyslexia: What are the boundaries? Data from Greek children. Research in Developmental Disabilities, 49–50, 339–353. Thordardottir, E., & Brandeker, M. (2013). The effect of bilingual exposure versus language impairment on nonword repetition and sentence imitation scores. Journal of Communication Disorders, 46, 1–16. Tsimpli, I. M. (2001). Interpretability and language development: A study of verbal and nominal features in normally developing and SLI Greek Children. Brain & Language, 77, 432–452. van der Lely, H. K. J. (1998). SLI in children: Movement, economy, and deficits in the computational-syntactic system. Language Acquisition, 7, 161–192. van der Lely, H. K. J. (2005). Domain-specific cognitive systems: Insight from Grammatical-SLI. Trends in Cognitive Sciences, 9, 53–59.

Downloaded from fla.sagepub.com by guest on June 10, 2016

294

First Language 36(3)

van der Lely, H. K. J., & Marshall, C. (2011). Grammatical-specific language impairment: A window onto domain specificity. In J. Guendouzi, F. Loncke, & M. J. Williams (Eds.), The handbook of psycholinguistic and cognitive processes: Perspectives in communication disorders (pp. 401–417). New York, NY: Psychology Press. van der Lely, H. K. J., & Ullman, M. (1996). The computation and representation of past-tense morphology in specifically language impaired and normally developing children. In A. Stringfellow, D. Cahana-Amitay, E. Hughes, & A. Zukowski (Eds.), Proceedings of the 20th Annual Boston University Conference on Language Development (Vol. 2, pp. 792–803). Somerville, MA: Cascadilla Press. Vandewalle, E., Boets, B., Ghesquière, P., & Zink, I. (2010). Who is at risk for dyslexia? Phonological processing in five-to seven-year-old Dutch-speaking children with SLI. Scientific Studies of Reading, 14, 58–84. Varlokosta, S. (2000a). Comparative conclusions on the acquisition of personal pronouns in typical development and SLI. In Proceedings of the 8th Symposium of the Panhellenic Association of Speech and Language Therapists (pp. 379–393). Athens, Greece: Ellinika Grammata. Varlokosta, S. (2000b). Paratirisis sto proto stadio tis glosikis anaptixis enos Ellinopoulou me Idiki Glosiki Diatarahi (Observations on the first stage of the linguistic development of a Greek SLI child). In Proceedings of the 8th Conference on Speech Therapy (pp. 160–174). Athens, Greece: Ellinika Grammata. Varlokosta, S. (2005). Eventivity, modality and temporal reference in child Greek. Advances in Greek Generative Syntax: In Honor of Dimitra Theophanopoulou-Kontou, 76, 217–240. Varlokosta, S., Vainikka, A., & Rohrbacher, B. (1998). Functional projections, markedness, and ‘root infinitives’ in early child Greek. Linguistic review, 15, 187–208. Verhagen, J., & Blom, E. (2014). Asymmetries in the acquisition of subject–verb agreement in Dutch: Evidence from comprehension and production. First Language, 34, 315–335. Vogindroukas, I., Protopapas, A., & Stavrakaki, S. (2009). Action Picture Test (Greek version of the Renfrew Action Picture Test). Chania, Greece: Glafki.

Downloaded from fla.sagepub.com by guest on June 10, 2016