Figure 2 - available via license: Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported
Content may be subject to copyright.
Source publication
Corpus linguistic methods can provide detailed and statistically robust information about how children's written language develops as they progress through their education. Such data can inform both models of written language development and curricular policies and practices. To this end, the current paper focuses on subordination as a key site of...
Contexts in source publication
Context 1
... initial analysis of subordinate clauses in the corpus showed adverbial clauses to be the most commonly-used type of subordination, with an average of .36 occurrences per main clause (see Figure 2). Finite adverbials are the most frequent subtype, with .25 occurrences per main clause, while non-finites appear, on average, .11 ...
Context 2
... contrast does not appear to be a feature of finite versuus non-finite clauses in general. Figure 2 showed that for many types of subordination (noun complements; prepositional objects; adjective complements, subjects), the non-finite is the more frequent form in child writing and our informal analyses of the development of non-adverbial clauses across year groups (not reported here) does not show finite clauses to be faster developing than non-finite clauses in general. ...
Citations
... Sentential relative clauses do have an antecedent, but the antecedent is the whole preceding clause, rather than a noun/pronoun. Biber et al., 2016;Biber et al., 2020;Durrant et al., 2020). Norwegian learners of English acquire -ing clauses from scratch in a very gradual fashion, whereas they use finite adnominal relative clauses earlier and much more frequently (Dirdal, 2021(Dirdal, , 2022. ...
This paper reports on an exploratory study of cross-linguistic influence in the acquisition of relative clauses by young Norwegian learners of English, comparing L1 Norwegian and L2 English material from the TRAWL (Tracking Written Learner Language) Corpus to L1 English material from the GiG (Growth in Grammar) Corpus. Previous reports of cross-linguistic influence in this domain have usually involved language pairs that have very different relativization strategies. This study investigates whether similarities between relative clause systems may lead to more subtle effects in the choice of relativizer, the type of head nominal, the syntactic function of the relativized item, the extent of relativization from embedded clauses and the use of relative clauses in special constructions such as existentials and clefts. Although the material is limited, the study found traces of the Norwegian system in the learners L2 English, signalling that this is an area worth further investigation. The learners struggled with the choice between who and which, but used that/zero in a very similar way to their L1 English peers. The L2 English group also had slightly higher frequencies of relative clauses belonging to existentials and clefts, and where the relativized item stemmed from a further embedded clause. These results are consistent with a usage-based theory of second language acquisition, where learners are assumed to transfer features of constructions from their L1 when they are similar enough for them to make a cross-linguistic identification.
... was granted through a formal request by the second author of this study for research purposes, which could be extended to the use of individuals connected to the same purpose. The complete GiG corpus, which had originally been divided into five subcorpora representing the grade levels year 2, 4, 6, 9 and 11, was rearranged for the purposes of the present study, excluding year 4 due to its limited number of texts (only 49) following the lead of two previous studies with the GiG corpus (Durrant and Brenchley, 2019;Durrant et al., 2020). The composition and features of the corpus used in this study can be seen in Table 2 below. ...
... Taking the genre of writing into account to see whether the use of CMs could vary depending on the types of writing, we analysed the corpus by categorizing texts under two over-arching labels, literary and non-literary as suggested by Durrant et al. (2020), which Table 11 and Fig. 2 below illustrate the results of the CMs used in each grade level in these two types of writing, which clearly indicates that literary texts included more CMs than those of non-literary texts. ...
... The second is Y9, where CMs are again used significantly less in the non-literary genre than in the literary writing. These findings appear to be in line with previous research that (i) registers, e.g., academic prose and conversation (Biber et al., 2000) and Spoken English, Academic Writing, Fiction, News Writing, and Other writing (Liu, 2008), (ii) disciplines, e.g., sciences and non-sciences (Peacock, 2010), and (iii) genres, e.g., literary and non-literary (Durrant et al., 2020) differ in terms of usage patterns and frequencies of various linguistic items, to which CMs in our study belong. ...
The use of contrast markers (CMs), one important type of connective creating semantic links both within and beyond the sentence (Robertson, 1968), is an explicit way of text cohesion (Halliday & Hasan, 1976) which develops gradually in childhood as logical reasoning advances (Piaget, 1928). This paper investigates the developmental trajectory of CMs in English across four grade levels to understand a particular aspect of language acquisition, that is ‘expressing various types of contrast’. We aim to reveal how frequently and diversely CMs are employed by children and report the differences between the overall use and that of the most frequent CMs across grade levels and types of texts. Through a frequency-based quantitative approach and contextual analyses, a set of 65 CMs constructed drawing on several existing taxonomies were searched in the Growth in Grammar (GiG) Corpus (Durrant & Brenchley, 2018), a corpus of school writing produced by children at schools in England from Year 2 to Year 11. The results show that the variety and the number of CMs increase across grade levels signalling a significant change as they get older. We also found that the frequencies of CMs differed significantly both across grade levels and genres of writing (literary vs. non-literary texts) in the corpus. In addition, it is noteworthy to evidence that the frequency of overall use of CMs significantly varied between Y2-Y6; Y6-Y9 and Y6-Y11, suggesting that Y6 (ages 10-11) stands in the middle of the V-shaped developmental curve. The findings also indicated that there are significant differences in the most frequent CMs (i.e., but, yet, and although) for each grade level.
... In the present study, we opted against using LOCNESS for reasons of comparability. However, the Growth-in-Grammar corpus (Durrant & Brenchley 2019;Durrant et al. 2020), comprising texts by British schoolchildren aged 8-16, may provide a good basis for comparison in future explorations. A reference corpus of comparable Norwegian L1 texts would offer an interesting opportunity to investigate crosslinguistic influence. ...
A pervasive finding in learner corpus research is that advanced EFL learners tend to overuse interactional features of writer/reader visibility (WRV) in their written academic texts, including first- and second-person pronouns, I think, modal adverbs, modal auxiliaries, and questions. Very little research has been done on younger learners, however. The present study is a mixed-methods investigation of WRV features in argumentative and expository genres in the TRAWL longitudinal corpus of learner texts from Norwegian lower secondary school. Comparisons are made with more advanced levels (undergraduate university students) using the Norwegian component of ICLE, ICLE-NO.
The results show that the TRAWL pupils use many WRV features in their writing, first-person reference being especially frequent (with I dominating). Compared to the advanced learners in ICLE-NO, the TRAWL learners overuse some, but not all, features. One explanation for the high frequency of WRV features in TRAWL is that the prompts – both argumentative and expository – often request a personal style. Some expository prompts and texts are more impersonal, but overall there is little distinction between the genres. The pedagogical implications are that instructors need to be more specific about genre requirements, and create more obligatory prompts that do not request a personal style.
... Investigators have employed various methods of measuring that proficiency, with different methods more or less suitable at different stages of development. For example, overall length of text rises sharply and then plateaus through the school years (Durrant et al., 2020), making it a less useful feature for tracking the ongoing writing development of adolescents. Instead, researchers tend to evaluate emerging skill with written language by increases in length of separate clauses or utterances (Scott & Windsor, 2000). ...
... A related skill that develops through adolescence involves the application of appropriate clause structures to various writing genres (Brimo & Hall-Mills, 2019;Nippold et al., 2008;Ravid & Tolchinsky, 2002). Skill with syntactic complexity in writing is not defined so much by a general ability to produce long clauses but rather by the metalinguistic insight needed to apply them appropriately across contexts (Durrant et al., 2020). For example, Berninger et al. (2011) showed evidence of increased use of subordinate clauses, particularly adverbial clauses, in narratives compared with nonnarratives in children's writing. ...
... Adverbs index more sophisticated and descriptive lexical choices. They also mark greater syntactic complexity and clausal density through adverbial clauses, the most common type of subordinate clause in corpus research on writing samples conducted in Durrant et al. (2020). ...
Purpose
In spite of improvements in language outcomes for children with hearing loss (HL) arising from cochlear implants (CIs), these children can falter when it comes to academic achievement, especially in higher grades. Given that writing becomes increasingly relevant to educational pursuits as children progress through school, this study explored the hypothesis that one challenge facing students with CIs may be written language.
Method
Participants were 98 eighth graders: 52 with normal hearing (NH) and 46 with severe-to-profound HL who used CIs. Oral and written narratives were elicited and analyzed for morphosyntactic complexity and global narrative features. Five additional measures were collected and analyzed as possible predictors of morphosyntactic complexity: Sentence Comprehension of Syntax, Grammaticality Judgment, Expressive Vocabulary, Forward Digit Span, and Phonological Awareness.
Results
For oral narratives, groups performed similarly on both morphosyntactic complexity and global narrative features; for written narratives, critical differences were observed. Compared with adolescents with NH, adolescents with CIs used fewer markers of morphosyntactic complexity and scored lower on several global narrative features in their written narratives. Adolescents with NH outperformed those with CIs on all potential predictor measures, except for Sentence Comprehension of Syntax. Moderately strong relationships were found between predictor variables and individual measures of morphosyntactic complexity, but no comprehensive pattern explained the results. Measures of morphosyntactic complexity and global narrative features were not well correlated, suggesting these measures are assessing separate underlying constructs.
Conclusions
Adolescents with CIs fail to show writing proficiency at high school entry equivalent to that of their peers with NH, which could constrain their academic achievement. Interventions for children with CIs need to target writing skills, and writing assessments should be incorporated into diagnostic assessments.
Supplemental Material
https://doi.org/10.23641/asha.17139059
... The rapid and remarkable changes child language undergoes before age five justify the amount of research for the earliest stages of acquisition, which is the framework underlying all the aforementioned studies. However, under the assumption that linguistic competence keeps growing during the school years as a result of explicit literacy instruction (Karmiloff-Smith, 1986;Kellogg, 2008;Durrant et al., 2020) and memory-related constraints (McCutchen, 2011), research on "later language acquisition" has gained increased attention prompted by the awareness that "becoming a native speaker is a rapid and highly efficient process but becoming a proficient speaker takes a long time" (Berman, 2004). Also in this scenario, which is the focus of our study, corpus-based approaches complemented with linguistically-informed indices (semi)-automatically extracted from text have started being applied to track the development of writing skills throughout the school years. ...
In this study we present a Natural Language Processing (NLP)-based stylometric approach for tracking the evolution of written language competence in Italian L1 learners. The approach relies on a wide set of linguistically motivated features capturing stylistic aspects of a text, which were extracted from students’ essays contained in CItA (Corpus Italiano di Apprendenti L1), the first longitudinal corpus of texts written by Italian L1 learners enrolled in the first and second year of lower secondary school. We address the problem of modeling written language development as a supervised classification task consisting in predicting the chronological order of essays written by the same student at different temporal spans. The promising results obtained in several classification scenarios allow us to conclude that it is possible to automatically model the highly relevant changes affecting written language evolution across time, as well as identifying which features are more predictive of this process. In the last part of the article, we focus the attention on the possible influence of background variables on language learning and we present preliminary results of a pilot study aiming at understanding how the observed developmental patterns are affected by information related to the school environment of the student.
... In this particular experiment, we made the decision not to use LOC NESS because we wanted to make sure that there was consistency across all of the factors that we were analyzing.On the other hand, future research could find it helpful to use the Growth-in-Grammar corpus, a collection of works created by students in the United Kingdom between the ages of 8 and 16. The government of Great Britain created this corpus(Durrant & Benchley, 2019;Durrant et al., 2020). Furthermore, it would be more fascinating to have the ability to investigate cross-linguistic impacts if there was a reference corpus of comparable works that had been authored in Norwegian as their L1 language. ...
According to the findings of one study that used a learner corpus, advanced students of English as a Foreign Language (EFL) frequently make errors in the usage of interactional WRV components in their written academic papers. I believe inquiries, first- and second-person pronouns, modal adverbs, and modal auxiliaries are all included in this category of characteristics. The findings of learner corpus research lead to this widely accepted and acknowledged conclusion. On the other hand, there hasn't been much study conducted on elementary and secondary school students. The primary objective of this research is to study WRV characteristics in argumentative and explanatory genres within the TRAWL longitudinal corpus of learner texts from Norwegian lower secondary schools using a combination of qualitative and quantitative research approaches. Expository, argumentative, and hybrid writing are all types of writing that fall under this category. ICLE-NO, the Norwegian adaptation of the exam, is utilized to compare levels that have progressed in their growth (undergraduate students at universities). According to the findings, students who take TRAWL are more likely to include WRV features in their writing, with allusions to first-person views being particularly widespread among these students' contributions (with I dominating). The students who are enrolled in TRAWL have less experience than those who are enrolled in ICLE-NO, and as a result, they commonly abuse a range of features. Nevertheless, this is not the case for every TRAWL student. The fact that the prompts, whether argumentative or explanatory, typically look for a personal style is one explanation for the high frequency of WRV in TRAWL. Other possible explanations include: The following are some more reasons that might be applicable: There are particular explanation prompts and messages that are less personal, but other than that, there is not much difference between the genres. Consequently, teachers must be more specific about the genres they want their students to generate and provide them with more necessary prompts without imposing any particular writing style on them. In addition, they will need to provide their students with more opportunities to write.
The topic of the present study is adverb-adjective combinations in narrative writing by lower secondary school pupils in Norway and the UK. The investigation is based on subsets of the TRAWL (Tracking Written Learner Language) and GiG (Growth in Grammar) corpora and thus compares English as a second language with first-language usage (EL2 and EL1). A number of differences were identified between the two writer groups. While adverb-adjective constructions, such as so happy, much better and really bad, were more frequent and widespread in EL2, they showed more variability in EL1 regarding syntax, semantics and lexical choice. In particular, the amplifying function of modifiers was more dominant in EL2 writing at the cost of other modifier functions. There was also a stronger concentration on a few highly frequent intensifiers in the EL2 than in the EL1 material.
Complex noun phrases (NP) are central to mature academic writing and often a focus of explicit teaching. The National Curriculum in England, for example, requires specific components of NP complexity to be taught at specific educational stages. However, the evidence base for such practices is unclear. Research on the emergence of NP components is both limited and dated. Moreover, some work has suggested that NP development is late-occurring and genre-specific, calling into question curricular guidance which specifies teaching from the earliest years and which makes no mention of genre. Analysing 240 texts written by children in England aged six to 16, this study shows that overall complexity develops at a roughly constant rate from primary school onwards. Increases are principally driven by postmodification, especially relative clauses and proposition phrases. By the end of their mandatory education, children make some use of genre distinctions evident in adult writing. However, there are also clear patterns of overuse and underuse of particular NP components. Key distinctive features are examined in context to understand the roles NP components play in writing development.
