Scholarly article on topic 'Development of Grammatical Aspect in Specific Language Impairment: Evidence from an Experimental Design with Video Stimuli'

Development of Grammatical Aspect in Specific Language Impairment: Evidence from an Experimental Design with Video Stimuli Academic research paper on "Psychology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Procedia Computer Science
OECD Field of science
Keywords
{Greek / "language disorders" / SLI / "semantics – morphology interface"}

Abstract of research paper on Psychology, author of scientific article — Katerina Konstantzou

Abstract Typical development of tense and aspect has been extensively studied across various languages. However, relevant research on language disorders has been generally restricted on tense morphology rather than aspect resulting in inconsistent findings. Specifically, some studies imply that acquisition of aspect in Specific Language Impairment (SLI) does not differentiate from typical development, whereas others report a deviant performance. These inconsistencies could be related to the limited number of languages tested or even to the different methodologies used across the studies. The present study examines the comprehension and production of grammatical aspect in Greek-speaking children with SLI. Thirty-six children participated in a combined comprehension – production task; twelve children with SLI (mean age: 6;3), twelve age-matched typically developing children (mean age: 6;3) and twelve language-matched typically developing children (mean age: 4;4). The originality of the study mainly lies on its methodology, since the task distributed to the children consisted of video stimuli instead of pictures, which are usually used in similar experimental designs. According to the findings, Greek-speaking children with SLI exhibited the same asymmetrical pattern reported for early stages of typical development. That is, perfective aspect is fully acquired, while interpretation and use of imperfective aspect seems to be problematic. Therefore, development of aspect in Greek SLI appears rather delayed than severely impaired.

Academic research paper on topic "Development of Grammatical Aspect in Specific Language Impairment: Evidence from an Experimental Design with Video Stimuli"

(8)

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Procedía Computer Science 65 (2015) 510 - 518

International Conference on Communication, Management and Information Technology (ICCMIT

Development of Grammatical Aspect in Specific Language Impairment: Evidence from an Experimental Design with Video

Stimuli

Katerina Konstantzoua*

aAKMI Metropolitan College, Sorou 74, Athens 12151, Greece

Abstract

Typical development of tense and aspect has been extensively studied across various languages. However, relevant research on language disorders has been generally restricted on tense morphology rather than aspect resulting in inconsistent findings. Specifically, some studies imply that acquisition of aspect in Specific Language Impairment (SLI) does not differentiate from typical development, whereas others report a deviant performance. These inconsistencies could be related to the limited number of languages tested or even to the different methodologies used across the studies. The present study examines the comprehension and production of grammatical aspect in Greek-speaking children with SLI. Thirty-six children participated in a combined comprehension - production task; twelve children with SLI (mean age: 6;3), twelve age-matched typically developing children (mean age: 6;3) and twelve language-matched typically developing children (mean age: 4;4). The originality of the study mainly lies on its methodology, since the task distributed to the children consisted of video stimuli instead of pictures, which are usually used in similar experimental designs. According to the findings, Greek-speaking children with SLI exhibited the same asymmetrical pattern reported for early stages of typical development. That is, perfective aspect is fully acquired, while interpretation and use of imperfective aspect seems to be problematic. Therefore, development of aspect in Greek SLI appears rather delayed than severely impaired.

© 2015TheAuthors.PublishedbyElsevier B.V.This is an open access article under the CC BY-NC-ND license

(http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of Universal Society for Applied Research

Keywords: Greek; language disorders; SLI; semantics - morphology interface

* Katerina Konstantzou. Tel.: +302106199891; fax: +302106199320. E-mail address: kkonstantzou@metropolitan.edu.gr

1877-0509 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of Universal Society for Applied Research

doi:10.1016/j.procs.2015.09.124

1. Introduction

Tense and grammatical aspect offer the two basic and complementary ways in which a situation can be related to the time line. In fact, tense locates the situation on the time line, whereas grammatical aspect expresses the speaker's perspective on a given situation with the basic distinction being between perfective and imperfective aspect. Perfective (PERF) aspect presents a complete event that takes place within certain time limits, e.g. Mary ate an apple. Imperfective (IMPERF) aspect reveals either a habitual or a continuous action, which is indefinite concerning time limits, e.g. Every day Mary is eating an apple or Mary was eating an apple [1] [2]. There is a cross-linguistic variety of how languages encode grammatical aspect with the most common ways being through verb morphology, particles, or aspectual adverbial phrases.

Research on the typical development (TD) of grammatical aspect has revealed conflicting results. Some studies report an early mastery of both aspectual distinctions [3] [4], while others found that grammatical aspect is not fully acquired before the age of 5;0 or even later [5] [6] [7]. Finally, a series of studies support that grammatical aspect develops following an asymmetrical pattern, that is, semantics of PERF aspect is acquired quite early, whereas interpretation and/or use of IMPERF aspect remains inadequate even at the age of 6;0 [8] [9] [10].

Specific Language Impairment (SLI) is a developmental disorder of language in the absence of hearing impairment, mental disability, motor-articulatory impairment, frank neurological impairment or psycho-emotional disturbance [11]. Nevertheless, the accurate specification of an SLI phenotype seems to be rather challenging, since children with SLI do not form a homogenous group. In fact, symptoms can vary across different ages affecting one or more linguistic levels. Moreover, precursors of the disorder may also vary across languages and it is particularly interesting that bilingual children with SLI do not necessarily exhibit the same deficits in both languages [12]. Despite the heterogeneity of the observed deficits, morphosyntactic problems can be seen as a hallmark of the disorder [13]. Indeed, a number of cross-linguistic studies confirm that morphosyntax is the level, which mainly is affected in SLI [14] [15] [16]. With respect to tense and aspect, many researchers have provided evidence that tense and agreement morphology are severely impaired in SLI [17] [18] [19]. On the other hand, research on the development of aspect and interface between morphology and semantics is rather restricted. Moreover, relevant findings are not always in line.

A study of 2003 tested the development of IMPERF aspect in English SLI [20]. English lacks the PERF - IMPERF distinction. Progressive aspect is marked with -ing, while habituality is usually expressed through present tense. On the other hand, PERF aspect is mainly associated with past tense due to pragmatic reasons. The researchers examined the use of third person singular -s and -ing as markers of habitual and ongoing events respectively. Overall, forty-five children participated in two elicited production tasks; fifteen children with SLI (range = 4;6 - 6;7, mean = 5;2), fifteen TD age-matched (range = 4;4 - 6;8, mean = 5;3), and fifteen TD language-matched controls (range = 2;8 - 4;11, mean = 3;6). The experimenter was acting out events with a puppet and then, the child had to either describe the event, which was happening at that time (progressive) or complete a sentence starting with "every day the puppet..." referring to the same event (habitual). Children with SLI did not differentiate compared to their controls, since they, actually, interpreted -ing as an ongoing and third person singular -s as a habitual marker. Based on this finding, the authors claim that grammatical aspect in English SLI is not vulnerable [20].

On the other hand, another study provided evidence that Cantonese-speaking children with SLI face severe difficulties in the use of grammatical aspect in past context [21]. The study examined the production of PERF and progressive grammatical aspect markers in Cantonese, which lacks grammatical tense. Grammatical aspect is coded through specific affixes, which generally are grammatically optional. However, there are syntactic contexts and pragmatic conditions, in which they are highly favored, if not obligatory. Fifteen children with SLI (range = 4;2 - 6;8, mean = 5;0), fifteen TD age-matched controls (range = 4;1 - 6;9, mean = 5;0), and fifteen TD language-matched controls (range = 2;11 - 3;6, mean = 3;2) participated in an elicitation task. Children had to describe past ongoing or complete events based on pictures from a book. The authors found that children with SLI performed worse compared to both control groups; hence, they concluded that the development of grammatical aspect appears deviant in Cantonese SLI, just like tense morphology appears deviant in other languages [21].

On the same path, a more recent study confirms that children with SLI face problems with production of grammatical aspect [22]. The study examined comprehension and production of PERF and IMPERF aspectual markers in past context by Hungarian-speaking children with SLI. In Hungarian, grammatical aspect is coded by the presence or absence of specific prefixes. Twenty-one children with SLI (range = 4;10 - 7;2, mean = 5;9), twenty-one TD age-matched controls (range = 4;8 - 7;3, mean = 5;8), and twenty-one TD language-matched controls (range = 3;3 - 6;6, mean = 4;10) participated in a combined comprehension - production task based on picture selection and description, respectively. According to the results, comprehension of both aspectual markers appears intact in SLI. On the contrary, their performance on production was less accurate compared to both control groups. Based on this finding, the authors claim that grammatical aspect appears deviant in Hungarian SLI, even though comprehension is intact [22].

Summarizing previous studies, it becomes evident that languages tested so far display differences in their aspectual systems, which may affect results. Moreover, the procedures used so far are mainly picture-based, which may also play a significant role in children's performance. Finally, the majority of the studies have not examined both comprehension and production so as to obtain a complete description of the course of acquisition. The present study aims to investigate comprehension and production of aspectual distinctions in past contexts by Greek-speaking children with SLI using an experimental design with video stimuli. Greek distinguishes between PERF and IMPERF aspect and encodes grammatical aspect through verb morphology [23]. Specifically, the majority of Greek verbs display at least two stems, namely a PERF and an IMPERF one. As a result, verbs obligatorily carry grammatical aspect, since they come in aspectual pairs [24].

The main research questions are the following:

• Are Greek-speaking children with SLI able to distinguish between PERF and IMPERF aspect?

• Is comprehension of aspectual markers vulnerable in Greek SLI?

• Is production of aspectual markers vulnerable in Greek SLI?

• Does the use of new technology (video stimuli) facilitate children's performance?

The present study is largely based on a previous study of 2013 [25]. Yet, it reports additional data from more children with SLI. The aforementioned study was the first one to examine the development of grammatical aspect in Greek SLI. The main finding was that development of aspectual distinctions in Greek SLI appears delayed rather than deviant, since it follows the asymmetrical pattern that has been reported for early stages of typical development [8] [9] [10].

2. Materials and Methods

2.1. Participants

Thirty-six monolingual Greek-speaking children participated in the study. They were divide into three groups; the SLI group, the TD age-matched control group, and the TD language-matched control group. Twelve children (8 boys, mean age (months) = 75, range = 59 - 94, SD = 9.24) comprised the SLI group. All children were recruited from the Community Mental Health Centre of Byron-Kessariani, 1st Department of Psychiatry, Medical School, University of Athens, Greece, and were diagnosed with a specific developmental language disorder (F.80, International Classification of Diseases, 10th Edition - ICD-10, 1990) by specialized staff. The selection criteria used were the same as reported in the relevant literature for SLI [11]. It should be mentioned that all children had been enrolled in a language intervention program for almost a year prior to their participation in the study. All children with SLI were individually matched to one TD child of the same chronological age. These twelve TD children (7 boys, mean age = 75, range = 59 - 93, SD = 9.05) comprised the age-matched control group (AM group). Each SLI child was also individually matched to one TD child based on the raw scores of the Greek adaptation of the Peabody Picture Vocabulary Test-Revised (PPVT-R) [26]. These twelve TD children (7 boys, mean age = 52, range = 43 - 73, SD =

9.65) served as the language-matched control group (LM group).

2.2. Materials and Procedures

All children participated in a combined comprehension - production experimental task designed within the COST Action A3 3 [27]. The innovation of the task lies on the use of video stimuli instead of pictures, which were traditionally used in similar tasks. Specifically, the materials consist of six short videos with a clown as the main character. The experimenter explains to the child that in every movie the clown has to perform six times a specific action (e.g., drawing) employing six different objects. However, there is a rule based on a popular children's game. That is, the clown can only move while music is playing and every time the music stops, he has to remain still until the music starts again. Then, he is allowed to move on to the next item - object. In half of the times, he succeeds to complete the action while the music is playing, so when the music stops, he freezes next to a complete project, e.g., a complete sun drawing (complete situation). In the other half of the times, he is still engaged with the project when the music stops, so he freezes next to an incomplete action, e.g., an incomplete star drawing (incomplete situation). Therefore, in each movie there are three complete and three incomplete events, which appear in pseudo-randomized order. The first four events test comprehension. That is, the experimenter asks the child to accept or reject the use of PERF and IMPERF aspect for complete and incomplete events. The music forms the background time period within which the child has to judge the aspectual value of the question, e.g., "While the music was playing, the clown was drawing/ drew + object". The last two situations always depict one complete and one incomplete situation and are used to test production. Specifically, the child has to complete the sentence "While the music was playing, the clown..." using the right aspectual type, e.g., was drawing/ drew + object. Overall, six different verbs were used; one for each movie. All verbs had regular inflection and were transitive forming a telic predicate.

With respect to the comprehension part, scoring was straightforward, since all answers were evaluated as correct or incorrect according to the target response. That is, IMPERF aspect was counted as correct for both complete and incomplete situations, since it does not entail completion. In contrast, PERF aspect carries completion and therefore, it should be rejected for incomplete events. Chance level was set at 50% taking into account that children had to choose between two possible answers (accept - yes or reject - no). With respect to the production part, the experimental set-up of the study (while-clause context) promoted the use of PERF as the only appropriate aspectual form for complete situations. On the other hand, IMPERF was the target aspect for the incomplete situations, as it focuses on a part of the event without providing information about completion. However, the scoring of the data was not as straightforward. That is, a variety of different answers was elicited, as children produced various responses. Table 1 presents children's answers and relevant coding. It should be mentioned, though, that production of PERF verb forms for incomplete events, was semantically incorrect, since PERF aspect entails completion.

2.3. Scoring

Table 1. Types of different answers that children produced for complete and incomplete situations

Complete situations

Answer type

Example

PERF aspect (target)

"He drew it"

IMPERF aspect

"He was drawing it"

IMPERF

Relevant expressions

"He finished it"

Incomplete situations

Answer type Example Code

IMPERF aspect (target) "He was drawing it" IMPERF

PERF aspect with negation "He didn't draw it" Negative-PERF

PERF aspect with half "He half-drew it" Half-PERF

PERF aspect "He drew it" PERF

Relevant expressions "He wasn't on time" Other

3. Results

3.1. Comprehension

The comprehension part targets to examine children's understanding of PERF - IMPERF distinction. Statistical comparisons (Binomial test) between participants' raw data and chance level showed that all groups, the SLI group included, performed above chance (p < .001). Thus, all children are capable to perceive the aspectual PERF - IMPERF distinction. Further statistical analyses were conducted in order to compare the three groups' performances. Specifically, a mixed-model analysis of variance (ANOVA) was used, with participant group (SLI, LM, AM) as a between-subjects factor and aspect (PERF - IMPERF) as within-subject factor reporting a significant group effect [F(2,33) = 6.881, p = .003]. Post-hoc testing (LSD) revealed that there was no significant difference between the SLI and the LM groups (p = .233), while both groups performed significantly worse compared to the AM group (SLI: p = .001; LM: p = .021).

Fig. 1 shows the mean percentages of target responses for each condition separately. Statistical analyses revealed a significant situation (complete - incomplete) effect (Wilcoxon, z = 2.209, p = .027), since all children performed better on complete rather than incomplete events. Specifically, all groups performed at ceiling for complete events accepting both aspect forms. With respect to incomplete events, though, they occasionally accepted PERF aspect. Nevertheless, all children who accepted PERF aspect for incomplete events, justified their acceptance providing explanations like "Yes, but he built half/ little of it". No statistical difference between the three groups was revealed for this condition (Kruskal-Wallis, %2(2, N = 36) = 1.648, p = .439).

As it is evident from Fig. 1, all children performed poorly on the incomplete - IMPERF combination. Specifically, the majority of the children tended to reject IMPERF aspect for incomplete events, even though IMPERF aspect is associated with ongoing events and does not carry completion. Statistical comparisons (Binomial test) between participants' raw data and chance level for this condition revealed that both the SLI and the LM groups performed below chance (p = .999). Moreover, no significant difference was reported between the two groups (p = .66). However, both groups performed worse compared to the AM group (SLI: p = .003; LM: p = .009). The AM group performed above chance (p = .005), but still their performance was low (65.27%).

Fig. 1 : Mean percentages of target answers for each condition

3.2. Production

The aim of the production part is to examine children's use of the aspectual forms when referring to both complete and incomplete events. Fig. 2a presents the mean percentage of each answer type produced for complete events and Fig. 2b for incomplete events (see also Table 1 for coding).

As it is evident from Fig. 2a, all children mainly used the target aspect, that is PERF verb forms, in order to describe complete events. Nevertheless, they occasionally produced IMPERF aspect as well. No statistical difference between the three groups was reported for production of either PERF (Kruskal-Wallis, %2(2, N = 36) = 3.049, p = .218) or IMPERF aspect (Kruskal-Wallis, %2(2, N = 36) = .754, p = .686). On the other hand, when children referred to incomplete situations (see Fig. 2b), they did not produce high rates of the target aspect form, namely IMPERF verb forms. A Mann-Whitney U test showed that the SLI and LM groups did not differentiate compared to each other (z = .876, p = .381). Nonetheless, both groups produced significantly lower percentages of IMPERF verb forms compared to the AM group (SLI: z = 2.267, p = .023; LM: z = 2.235, p = .025).

On the other hand, as it is evident from Fig. 2b, both the SLI and the LM groups produced high rates of PERF verb forms with negation. There was no significant difference between the two groups (Mann-Whitney U test, z = 1.816, p = .069). However, both groups produced this answer type significantly more often compared to the AM group, who used it less (Mann-Whitney U test, SLI: z = 2.346, p = .019; LM: z = 3.134, p = .002). An additional answer type that should be further discussed is the use of PERF verb forms without either negation or half. As it has already been mentioned, the production of PERF aspect for incomplete events is semantically incorrect, since it entails completion. Yet, the SLI group as well as the younger controls used PERF verb forms to describe incomplete events. A further analysis revealed that three out of twelve children with SLI and two out of twelve TD children used PERF aspect for incomplete events. Moreover, four of them used PERF aspect only once referring to the same complete situation (opening of a present). It should be mentioned that there was no statistical difference between the two groups (MannWhitney U test, z = .572, p = .568).

SU giop LV groMP ¡14 grrxf

3B*_0rotip

5U grap LV grajp Ul '1'M-3g*_group

BOter ■fE«F □Hil^ERF

'MPESF

Fig. 2: (a) percentages of each answer type produced for complete events; (b) percentages of each answer type produced for incomplete events

4. Discussion

Based on the data of the present study, Greek-speaking children with SLI are able to distinguish between PERF and IMPERF aspect. Furthermore, they seem to have fully acquired the semantic entailment of the PERF aspect, namely completion. Indeed, they rejected it for incomplete events and produced it to a great extend when referring to complete situations. When they occasionally accepted PERF aspect for incomplete events, they provided explanations like "Yes, but he drew half of it". In fact, TD children also sometimes accepted the aforementioned combination providing similar explanations. With respect to production, children with SLI incorrectly used PERF verb forms in order to refer to incomplete situations. However, a further analysis revealed that only three out of twelve children produced it and actually, two of them used it just once for a particular item; describing the incomplete opening of a present. This specific item could be confusing, since children may have understood the unwrapping of the present as the actual opening (while the clown is opening the present, he is actually unwrapping the paper). It is especially interesting that two younger TD controls also used PERF verb forms to refer to this particular item providing further evidence that this pattern cannot be considered as deviant. The main difficulty that was reported for the SLI group was the interpretation and use of IMPERF aspect. Specifically, they incorrectly rejected it to a high extend for incomplete events and they did not use it extensively in order to describe incomplete events. Consequently, Greek-speaking children with SLI exhibit an asymmetrical developmental pattern. That is, PERF aspect seems to be fully acquired, while interpretation and use of IMPERF aspect is problematic.

Nevertheless, that was the actual pattern observed for TD children as well. In fact, both control groups exhibited the same error - type; target knowledge of PERF aspect vs. non-target knowledge of IMPERF aspect. Specifically, both groups performed at ceiling for PERF aspect (comprehension - production). On the other hand, they had problems with IMPERF aspect, even though the older controls performed better compared to the younger controls. Altogether, the present data indicate that development of aspectual distinctions in Greek SLI is not deviant. Instead, it seems rather delayed, as it follows the same asymmetrical error pattern observed in earlier stages of typical development [8] [9] [10].

The present study confirms the study of 2013 providing further evidence that the development of aspect in Greek SLI is late and not vulnerable [25]. This finding contradicts previous studies that claim that production of aspect is deviant in SLI [21] [22]. As it has already been discussed in the study of 2013, these inconsistencies could be attributed to specific-language characteristics or even methodological issues [25]. As a matter of fact, materials of the present study consist of mini-videos, which succeed in accurately depicting time-related qualities like aspect. Specifically, there is a background time event (music playing), which functions as the time frame within which the use of each aspect type is evaluated. On the other hand, previous studies are mainly based on pictures asking from children either to produce isolated sentences or to pick the picture that is related with the sentence they hear.

Concluding, the present study examined comprehension and production of aspect in Greek SLI using an experimental design with video stimuli instead of picture-based tasks. Altogether, the present data indicate that development of aspect in Greek SLI is rather late than deviant contradicting previous studies [21] [22]. Nevertheless, given the limited number of relevant studies as well as the inconsistent findings reported across different languages, it becomes clear that more relevant studies should be contacted before we reach firm conclusions regarding the deficiency of grammatical aspect in SLI.

Acknowledgements

This research has been co-financed by the European Union (European Social Fund - ESF) and Greek national funds through the Operational Program "Educational and Lifelong Learning" of the National Strategic Reference Framework (NSRF) - Research Funding Program: THALES - UOA - "COGnitive MEchanisms in the Perception, Representation, and Organization of Knowledge (COGMEK)", P.I.: Stella Vosniadou. The author would like to thank the Community Mental Health Centre of Byron-Kessariani, 1st Department of Psychiatry, Medical School, National and Kapodistrian University of Athens, Athens, Greece, where the assessment of the children with SLI took place. The author is also grateful to the teachers of the daycares, kindergartens, and the elementary school that participated in the study: Aerostato, Palataki, 110th Kindergarten of Athens, 131st Kindergarten of Athens, 106th Elementary School of Goudi. Permission to assess the children at the kindergartens and at the elementary school was granted by the Ministry of Education.

References

[1] Comrie B. Tense. UK: Cambridge University Press; 1985.

[2] Comrie B. Aspect. UK: Cambridge University Press; 1976.

[3] Bar-Shalom E. Tense and aspect in early child Russian. Language Acquisition 2002; 10:321-327.

[4] Weist RM. Prefix versus suffix information processing in the comprehension of tense and aspect. Journal of Child Language 1983; 10:85-96.

[5] Hodgson M. The acquisition of Spanish perfective aspect: a study on children's production and comprehension. ZAS Papers in Linguistics 2003;29:105-117.

[6] Matsuo A. Young children's understanding of ongoing vs. completion in imperfective and perfective participle. Linguistics 2009;47(3):743-757.

[7] Wagner L. Understanding completion entailments in the absence of agency cues. Journal of Child Language 2002;29:109-125.

[8] Kazanina N, Phillips C. Russian children's knowledge of aspectual distinctions. In: Beachley B, Brown A, Conlin F, editors. Proceedings of the 27th Annual Boston University Conference on Language Development. Somerville, MA: Cascadilla Press; 2003. p. 390-401.

[9] Konstantzou K. Acquisition of tense and aspect: Evidence from production and comprehension tasks. Unpublished PhD Thesis. University of Athens; 2014.

[10] van Hout A. Imperfect imperfectives: on the acquisition of aspect in Polish. In: Kempchinky P, Slabakova R, editors. Aspectual inquiries. Dordrecht: Springer; 2005. p. 317-344.

[11] Leonard L. Children with specific language impairment. Cambridge, MA: MIT Press; 1998.

[12] Paradis J, Crago M, Genesee F. Domain-General versus Domain-Specific Accounts of Specific Language Impairment: Evidence from Bilingual Children's Acquisition of Object Pronouns. Language Acquisition 2006;13(1):33-62.

[13] Bishop DVM. Grammatical errors in Specific Language Impairment. Competence or performance limitations? Applied Psycholinguistics1994;15(4):507-550.

[14] Hamann C, Ohayon S, Dube S, Frauenfelder UH, Rizzi L, Starke M, Zesiger P. Aspects of grammatical development in young French children with SLI. Developmental Science 2003;6:151-158.

[15] Friedmann N, Novogrodsky R. The acquisition of relative clause comprehension in Hebrew: A study of SLI and normal development. Journal of Child Language 2004;31:661-681.

[16] van der Lely HKJ, Christian V. Lexical word formation in children with grammatical SLI: a grammar-specific versus an input-processing deficit? Cognition 2000;75:33-63.

[17] Jakubowicz C, Nash L, van der Velde M. Inflection and past tense morphology in French specific language impairment. In: Greenhill A, Littlefield H, Tano C, editors. Proceedings of the 23rd Annual Boston University Conference on Language Development. Somerville, MA: Cascadilla Press; 1999. p.289-300.

[18] Mastropavlou M. Morphophonological salience as a compensatory means for deficits in the acquisition of past tense in SLI. Journal of Communication Disorders 2010;43:175-198.

[19] Rice ML, Wexler K. Toward tense as a clinical marker of specific language impairment in English-speaking children. Journal oof Speech, Language, and Hearing Research 1996;39:1239-1257.

[20] Leonard L, Deevy P, Miller CA, Charest M, Kurtz R, Rauf L. The use of grammatical morphemes reflecting aspect and modality by children with specific language impairment. Journal ofChild Language 2003;30:769-795.

[21] Fletcher P, Leonard L, Stokes S, Wong AMY. The expression of aspect in Cantonese-speaking children with specific language impairment.

Journal of Speech, Language, and Hearing Research 2005;48:621-634.

[22] Leonard L, Lukacs A, Kas B. Tense and aspect in childhood language impairment: Contributions from Hungarian. Applied Psycholinguistics 2012;33:305-328.

[23] Holton D, Mackridge P, Philippaki-Warburton I. Grammar of Modern Greek. Athens: Patakis Publications; 2004.

[24] Mozer A. Aspect and tense in the history of Greek. 2nd ed. ParousiaMonograph Series'; 79. Athens: National and Kapodistrian University of Athens; 2009.

[25] Konstantzou K, van Hout A, Varlokosta S, Vlassopoulos M. Perfective - Imperfective: Development of aspectual distinctions in Greek Specific Language Impairment. Linguistic Variation 2013;13(2):187-216.

[26] Simos PG, Sideridis GD, Protopapas A, Mouzaki A. Psychometric characteristics of a receptive vocabulary test for Greek elementary students. Assessment for Effective Intervention 2011; 37(1):34-49.

[27] COST Action A33 http://www.cost.eu/COST_Actions/isch/Actions/A33