Speakers
Description
Introduction The research approach to semantic development in first language acquisition (FLA) remains predominantly enclosed in traditional lexicographic terms and notions (usually simplified). This viewpoint doesn’t adequately span the lexical system’s complexity and fails to present the mechanisms and processes involved in its development. Since FrameNet possesses standardized methods and procedures for the sufficient description of lexical system, we suggest it as a research method in studies of semantic development. Research subject and goal We explore FrameNet as a tool for detecting and tracking changes occurring during semantic development in FLA. The subject of the analysis is a Serbian LU “uspomena” and its usage in essays of elementary school students aged 9 to 15. This unit suits LU “memory.n“, frame: Remembering_experience in FrameNet. According to the Dictionary of the Serbian Language (Nikolić, 2011), the lexeme “uspomena” consists of two lexical units: 1. ‘a memory of someone or something, an impression of remembering someone or something’; 2. ‘an item that reminds of someone or something, which is kept in memory.’’ Using FrameNet, we also explore whether these LUs are split in the linguistic production of elementary school students in Serbia. 59Language resources Instances of target word usage were excerpted from four different corpora: (1) Developmental Corpus of Serbian Written Language (RAKOPS), (2) Referent Corpus of Contemporary Serbian (SrpKor2013 and SrpKor2021), (3) Serbian Web Corpus (SrWac), and (4) SrpELTeC (Serbian part of the European Literary Text Collection). Method Examples of target word usage extracted from corpora have been manually annotated for FEs and PTs by applying methods and procedures described in Ruppenhofer et al. (2016). The first step involved establishing FrameNet for the Serbian language and building its support.1 Next, FEs and PTs were established according to the target word usage in both referent corpora, web corpus, and ElTeC. This way, the reference point for comparing the usage of analyzed LU in school-age children’s written production was provided. Afterward, extracted LU instances from the RAKOPS were annotated for FEs and PTs. Lastly, established FEs and PTs were compared within RAKOPS, according to students’ grade level, and between RAKOPS and referent corpora. Some results Differences in FEs and PTs of analyzed LU between the usage in the RAKOPS and referent corpora were determined. A core element STATE isn’t confirmed in examples excerpted from the RAKOPS, while three FEs from another LU (“recollection.n”) are found (grey zone in Table 1). This shows that the demarcation line between the two senses of the lemma “uspomena” isn’t fully established, and that sense blending in part of peripheral and extra-thematic FEs occurs at this developmental stage. Also, differences in FEs and PTs according to students’ grade level were determined (Table 1). 1 First attempts are described in Marković et al. (2021). 6061 Table 1: FEs and PTs of non-standard usage examples in RAKOPS according to grade level Grade LU “uspomena,n. ” (Eng. memory,n.) FE and PT Cog- nizer Exp Imp Salient_ ent Context Man- ner Content Place Exp lan 3 CNI Kao-što. Sub AJP PP[u].Loc Kada. Sub Jer. Sub Np.Nom Srel Poss.Det PP[na].Dep AVP 4 Poss.Dat PP[u].Loc CNI AJP PP[u].Loc Poss.Dat Srel PP[u].Loc Poss.Nom PP[na].Loc Np.Nom AVP Poss.Dat Poss.Nom PP[u].Loc 5 Poss.Nom AVP CNI PP[u].Loc CNI AJP PP[u].Loc Poss.Det AVP Poss.Det AJP Srel PP[u].Loc 6 Poss.Dat PP[iz].Gen AVP Poss.Dat PP[u].Loc CNI AJP AVP CNI Kao-npr. NP AJP AVP CNI AJP CNI AVP / PP[u].Loc Poss.Dat AVP 7 CNI AVP Poss.Dat AVP / NP.nom PP[u].Loc CNI PP[od]. Gen. AJP 8 Poss.Dat AVP CNI PP[sa]. Ins. AJP CNI AVP CNI PP[iz].GenConclusion In the traditional approach, determined differences in the usage of the analyzed LU would be considered “lexical errors.” Nonetheless, the application of FrameNet enabled insightful findings that reveal the course of meaning development and sense splitting rather than a vacuous deviation from the standard usage.
