Bempen or Bemben: Differences Between Children At-Risk of Dyslexia and Children With SLI on a Morpho-Phonological Task

Abstract

This study assesses morpho-phonological alternation in plural formation by 5-year-old Dutch children with a familial risk of dyslexia, children with specific language impairment (SLI), and typically developing children. The morpho-phonological process investigated is the voicing alternation in Dutch singular–plural pairs such as bed [t] ‘bed’ ~ bedden [d] ‘beds’. Results on a plural formation task showed that the at-risk group, like the control group, produced more correct plurals in real words than the SLI group, and was sensitive to lexical frequency. However, the at-risk group patterned with the SLI group on nonword stimuli, repeating fewer correct singulars and producing fewer plurals. The SLI group's performance was further characterized by production difficulties affecting clusters, difficulties with morphological inflection, and possibly reduced sensitivity to lexical frequency and distributional patterns. These findings support the phonological deficit hypothesis of dyslexia and suggest that dyslexia and SLI should be considered separate disorders.

Notes

¹Children with a familial risk of dyslexia have at least one dyslexic parent.

²Note that the current study focuses on plosives, although fricatives (/f~v/ and /s~z/) undergo the same process in Dutch.

³Orthographic representations are presented in italics. The phonetic transcription is provided in brackets, indicating whether the plosive is voiceless [t, p] or voiced [d, b].

⁴For the −en [n] suffix, pronunciation of [n] is optional in standard Dutch.

⁵In accordance with Dutch orthography, the long vowel in [debn] is represented as a single grapheme.

⁶The higher number of −s plurals in the SLI group was due to one child, who produced 11 −s plurals and 4 −en plurals for the 15 words and 11 −s plurals and one null marking error for the 12 nonwords.

⁷Such a correlation was not found for non-alternating plurals, indicating that alternating plurals are stored as irregulars.

⁸We thank an anonymous reviewer for pointing this out. A paired samples t test comparing the proportion voicing alternations for alternating words and nonalternating words statistically confirms that the children with SLI produced more alternations for alternating words, t(23) = 3.12, p = .005.

⁹In some instances, stem changes co-occurred with an incorrect singular form (e.g., book~boke [bok] for boop, flat~flate [flt] for flant, ben~bene [bεn] for bemp), but in others only the plural forms were incorrect, mostly due to cluster reduction (e.g., kimp~kime [k1m], flant~flate [flt]).

¹⁰Paired samples t tests per group compared the difference between proportion plural realization for words and nonwords. There was a significant difference for the at-risk group only, t(57) = 3.92, p = .00, although the difference approaches significance for the SLI group, t(23) = 1.99, p = .058. The results of the control group are not different for words and nonwords (p = .36).