Cynthia Kilpatrick

Analysing frequency and temporal reliability of children’s morphosyntactic production in spontaneous language samples of varying lengths

Spontaneous language samples are a useful tool in studying the nature of morphosyntactic production in both clinical and academic settings. However, little work has been done on typical frequency of occurrence of morpheme production and even less on the degree of temporal reliability of morpheme frequency between samples. This study compares two language samples recorded within a week of each other under similar conditions in a test–retest paradigm for each of 23 typically developing children aged 2;6–3;6 to examine the production of the following morphemes associated with specific language impairment: contracted auxiliary, copula, do as an uncontracted auxiliary, be as an uncontracted auxiliary, third person singular and past tense. Also analyzed were all utterances containing more than one verb. Frequency and reliability of four different sample lengths ranging from 50 to 200 utterances were tested. Results showed generally low degrees of reliability at 50 and 100 utterances, with only limited improvement for most targets at 200 utterances. Clinical implications of these results are discussed.

The reliability of morphological analyses in language samples

It is currently unclear to what extent a spontaneous language sample of a given number of utterances is representative of a child’s ability in morphology and syntax. This lack of information about the regularity of children’s linguistic productions and the reliability of spontaneous language samples have serious implications for language testing based upon natural language. This study investigates the reliability of children’s spontaneous language samples by using a test-retest procedure to examine repeated samples of various lengths (50, 100, 150, and 200 utterances) in regard to morpheme production in 23 typically developing children aged 2;6 to 3;6. Analyses indicate that out of five morphosyntactic categories studied, one of these (the contracted auxiliary) achieves an ICC for absolute agreement over .6 using 100 utterances while most others (past tense, third-person singular and the uncontracted ‘be’ in an auxiliary form) fail to reach a correlation above .52 even when samples of 200 utterances are compared. The study indicates that (1) 200-utterance samples did not provide a significantly greater degree of reliability than 100 utterance samples; (2) several structures that children were able to produce did not show up in a 200-utterance sample; and (3) earlier acquired morphemes were not used more reliably than more recently acquired items. The notion of reliability and its importance in the area of spontaneous language samples and language testing are also discussed.

Intrusive vowels in Cruceno Spanish

Intrusive vowels are short vowels appearing within consonant clusters that are not treated phonologically like full vowels. Though reported in many languages, few quantititative studies have examined these vowels. Here, an acoustic study of intrusive vowels in obstruent+tap clusters in the Spanish of Santa Cruz, Bolivia is reported. The results support earlier studies in that the intrusive vowel quality resembles that of the following nucleic vowel, rather than a neutral vowel, and intrusive vowels are significantly longer in clusters with voiced obstruents than in those with voiceless ones. However, results do not fully support all previous phonetic descriptions and related theoretical assumptions. In particular, a significant difference is not found in intrusive vowel length based on variables such as place of articulation of the obstruent, quality of the nucleic vowel, or placement of stress in relation to the cluster. In addition, where previous work finds no significance for position in the word, the...

Child-directed speech: An examination of frequency and reliability in spontaneous language samples.

Abstract Little research exists regarding language reliability, meaning that it is unknown to what degree language samples represent an individual’s typical linguistic production. While few studies exist regarding child language reliability, fewer exist for adults. This study uses a test–retest procedure to examine frequency and reliability of morphosyntactic production in the language of 17 mothers interacting with their children. To examine reliability of different sample lengths, counts and reliability are calculated at 50, 100, 150 and 200 utterances. Results are compared to those of children (2;6–3;6) involved in a larger study and reported in previous work. Results show that (1) frequency counts were higher for adults than children, (2) reliability of most elements was markedly lower in adult language, and (3) relative frequency of items in the child language samples mirrored those in adult samples.

Second language influence on perception of first language phonotactics.

The results of a speech perception experiment are reported, examining the responses of speakers of the same L1 (Spanish) with varying experience in the L2 (English). Three groups of speakers participated in a wordlikeness rating task in Spanish, in which nonce words were presented in sets that varied as to whether the final consonant of the word was phonotactically legal or illegal: (1) bilinguals who learned both English and Spanish before the age of 5, (2) L1 Spanish speakers that learned English after the age of 12, (3) monolingual speakers without communicative exposure to a second language. All groups differed significantly in the strength of their overall ratings, even though all speakers were technically native speakers of Spanish. In addition, for both the bilingual and second language speakers, ratings for legal vs illegal codas were significantly different. For monolingual speakers, on the other hand, ratings for legal and illegal codas were similar, but response times for these two conditions w...

On the perception of epenthetic stops in American English

This study examines the perception of epenthetic stops in American English. Stimuli were created from prince, prints, mince, mints, quince, and quints by removing all traces of /t/ and splicing in 0–72 ms of silence, in 12‐ms steps, with or without a following burst. Subjects saw the minimal pairs on screen and selected the word they heard. It was hypothesized that stimuli with bursts and longer closure would result in more t responses (prince identified as prints) and that frequent words (prince/prints) would be more difficult to distinguish than infrequent words (quince/quints), as our production results suggest that frequent pairs are more likely to be produced similarly. Results from 19 subjects show shorter response times with longer silence intervals, but no effect of burst or stimulus identity. Similarly, stimuli with bursts were not identified as nts words more frequently than those without. Generally, stimuli originating from nts words were more likely to be identified as such if they contained a...