Analysis of maternal morphemes input provided to children using cochlear implants
1 like398 views
This study analyzes maternal language input to children with significant hearing loss using cochlear implants, revealing that parental mean length of utterance in morphemes (mlum) and words (mluw) tends to shorten compared to that used with hearing children. The research suggests correlations between maternal input length and the child's age and duration post-cochlear implantation, indicating a lack of exposure to essential morphemes. The findings highlight the need for clinicians to guide parents in providing enriched language input to enhance language acquisition in their children.
Analysis of maternal morphemes input provided to children using cochlear implants
- 1. Analysis of maternal morphemes: Input provided to children using cochlear implants Shani Dettman1,2, Melina Ramp1, Colleen Holt1, Richard Dowell1,2,3, Denise Courtenay3 The University of Melbourne, Department of Otolaryngology, 2 The HEARing Cooperative Research Centre, 3Royal Victorian Eye and Ear Hospital Aims of the Study. Very few studies have attempted to document the full range of linguistic expressions and constructions that children with normal hearing and children with significant hearing loss have exposure to in their daily lives. This study investigated the length of input that parents used when talking to their children with significant hearing loss, which enabled comparisons to be made with existing data for hearing children. It was hypothesized that the length of MLU would be fine-tuned to the child’s chronological age. Secondly, this study provided a detailed description of the types and prevalence of morphemes that were embedded in the parental language directed to the child. The nature and prevalence of linguistic forms provided by the mother is of considerable interest to clinicians working with children who have significant hearing loss. Clinicians currently guide and coach parents to provide the best possible language input and language models to the child, but understanding whether parents should be encouraged to use single words, whole phrases, or more enriched discourse input is unproven at this time. …to understand the nature of the relationship between maternal input ‘IN’ and the child’s language acquisition and ‘OUTPUT’ MLU Results The average maternal MLUm and MLUw was 3.51 and 3.02 respectively, and ranged from 1.26 to 8.74 for morphemes and 1.24 to 6.94 for words. Overall, parents did increase the length of their MLUm (red triangles) and MLUw (black squares) in line with the child’s age (Figure 2). There was a significant correlation between MLUm and age at sample (r=0.501, p<0.001) and duration post CI (r=0.664, p<0.001) and a significant correlation between MLUw and age at sample (r=0.509, p<0.001) and duration post CI (r=0.544, p<0.001). 9 Maternal MLU Background to this study For young hearing children who are learning language, the literature suggests that parents use modified linguistic input such as shorter utterances (Phillips, 1973; Snow, 1972), fewer broken or run on sentences, less complex but well formed grammatically correct sentences, and approximately 50% of utterances are single words or short declaratives (Owens, 1992, p. 233). These simplifications in the parent’s input appear to support vocabulary development and the acquisition of receptive and expressive language. Cross (1977) did not find that parental input was always fine-tuned to the child’s linguistic abilities and demonstrated low correlations between some variables under study. Overall, however, results suggested a relationship between parental discourse, referential and syntactic features with the child’s receptive skills, ability to get the message across, age, vocabulary (type-token ratio), and measures of utterance length (longest utterance and mean). Thus, the parents of the infants with normal hearing in Cross’ (1977) study appeared to be altering many facets of their input at different times. Parental MLU appears to be a good match to the child’s linguistic abilities and varies according to the language used and the age of the listener. The MLU results from a range of studies investigating input to children with normal hearing and normal language are charted against the child participants’ chronological age in Figure 1. Results suggest a longer maternal input length during the neonatal period (Penman et al., 1983; Phillips, 1973) while the child has the opportunity to listen to his/her native language. Subsequent to this, we see a low or falling maternal MLU in the child’s first year of life prior to the emergence of the child’s first words (Murray, Johnson & Peters, 1990), followed by a gradual increase in MLU with each child chronological year (Cramblit & Siegel, 1977; Cross, 1977; Kaye, 1980; Lasky & Klopp, 1982; Lord, 1975, cited in Owens, 1992; Newport, Gleitman & Gleitman, 1977; Seewald & Brackett, 1984). 6 MLU 3 0 1 2 3 4 5 6 7 8 Chronological Age (years) Figure 2. Parental MLU in morphemes (red triangles) and MLU in words (black squares) N=20 children with significant hearing loss using cochlear implants. There was a wide variation in maternal MLUm and MLUw to children with significant hearing loss (Figure 2) in comparison to the averages for parental input to children with normal hearing (Figure 1). Some mothers persisted with shorter MLU to their 3 and 4 year olds, as if they were speaking to toddlers. Importantly, there was an absence of longer maternal input early in the child’s development. Some parents reduced input to single words in a ‘teaching’ style of interaction. The authors suggest that parents were unaware of the importance of exposing the child to the full rich character of the language or were confused whether to fine-tune their MLU to the child’s hearing age, chronological age or linguistic ability. Seewald and Bracket (1991) suggested that parents of children with hearing loss may believe the common and naïve view of language learning; that it occurs in a ‘bottom-up’ rather than ‘top down’ fashion. Parents who perceive themselves in this teaching role may reflect this in their style of interacting, and may provide far more language at the word level rather than the phrase level. We observed (Figure 3, below) that children only had the opportunity to hear numerous examples of bound morphemes (20 or more bound morphemes in a 50 utterance sample) once the parental MLUm exceeded 3 or 4. Bound Morpheme Types The morphemes found in the present study that were also described by Brown (1973) are on the left of Figure 4, below (e.g. -ing, plural -s, irregular past tense, possessive –‘s, regular past tense -ed). Other bound morphemes, not described by Brown, continue from left to right in no particular order. The following were frequently heard by the children; • contracted copula BE e.g. that’s, he’s • contracted modal auxiliary will, e.g. she’ll • regular plural -s e.g. cats, • present progressive Ving form e.g. jumping In contrast, there were infrequent opportunities for the children to hear; • irregular plurals e.g. mouse/mice • comparatives e.g. big-bigger • superlatives e.g. biggest 100 80 60 40 20 0 0 2 3 4 6 8 10 Maternal MLU morphemes Figure 3. Raw score difference between bound morpheme and word total. child first words 0 25 8 20 Chronological Age (years) dettmans@unimelb.edu.au www.hearingcrc.org creating sound value unanalysed diminutive other prefix/suffix quantifier catenative intensifier preposition (compound) suffix - superlative - est, st -er - comparative -er, r -er - person's name/job adverb -ly negative - contracted contracted us 0 adjective -y, ly 5 irregular plural Twenty children who received a cochlear implant before 4 years of age, were recorded in play for over 10 minutes with their mothers pre- and post-implant (mean age at sample 2.74 yrs, range 1.04 to 7.70, SD 1.10). All maternal input was orthographically transcribed and analysed. Previous literature had used Browns (1973) 14 morphemes which is a simplified list of just some of the morphemes used in everyday speech. In order to complete this analysis, the authors developed two tools; a definition of the types of morphemes in Australian English and an alphabetical glossary to aid coding. The mean length of utterance in morphemes (MLUm) and words (MLUw) from 50 consecutive utterances, from 175 samples was calculated. Individual bound morphemes were defined and totalled for each sample. The raw score total of each bound morpheme type was expressed as a percentage of all bound morphemes to derive the prevalence. 10 yours , his, hers, theirs, ours Methods and Materials 15 -en Figure 1. Parental MLU to children with normal hearing. Each black box represents mean MLU results. % irregular 3rd person 7 contracted aux 6 contracted copula 5 regular past -ed 4 regular 3rd person -s 3 irregular past 2 possessive -s 1 -ing 0 regular plural -s Speaker's MLU 9 6 0 R a w S c o r e - d iffe r e n c e b e tw e e n w o r d a n d b o u n d m o rp h e m e s c o re s One measure of linguistic complexity is the mean length of utterances in ‘morphemes’ (MLUm). As the child learns language, he/she comes to understand that some words are made up of smaller functional parts. These smaller units of meaning are called morphemes and include: unbound morphemes (we refer to as words) consisting of sole morphemes (simple content words) such as ball, walk, car, function and social words such as no, a, to, you, and bound morphemes such as inflectional and derivational bound morphemes e.g. prefixes/suffixes or grammatical markers such as im-, dis-, possessive –s, and past tense –ed. Brown (1973) suggested that a child acquires these meaningful units in a typical order and there is some debate in the literature whether their order of acquisition is related to the parental input frequency, parental ‘rough-’ or ‘fine-’ tuning or due to other factors. Deafness Foundation compound word 1 Figure 4. Prevalence of each bound morpheme type expressed as a percentage of all bound morphemes, across 175 samples from N=20 children. Conclusions A social interactionist account of language acquisition would suggest that maternal language input has important facilitatory effects on child language development. This study has determined that for some cases, maternal MLU to children with significant hearing loss may be shorter than MLU in speech directed to hearing children. Although mothers in this study did increase MLU over time in line with the child’s chronological development, important exposure to particular morphemes appeared to be lacking. Infrequently heard bound morphemes are the same morphemes for which children with significant hearing loss typically demonstrate poor understanding and expression, suggesting that specific training is required. Acknowledgements to the patients, speech pathologists, audiologists , surgeons and administrative staff at the Cochlear Implant Clinic, RVEEH, Melbourne, Australia.