CPCSLD: A lexical database of Chinese preschool children's spoken words

To facilitate research into the language development of preschool children, this article presents the Chinese Preschool Children's Spoken Lexical Database (CPCSLD), which is a lexical database built from a corpus of spontaneous speech production by 648 Chinese children aged between 3 and 6 years. The corpus comprises 1,199,851 word tokens, which include 21,372 unique words, 1,147 unique tonal syllables, and 400 unique atonal syllables. CPCSLD provides multiple distributional characteristics of both word-level and syllable-level information, including word frequency, token frequency, word length, word syntactic categories, tonal syllable frequency, and atonal syllable frequency, for words in the entire corpus as well as in the three grade-level sub-corpora (K1, K2, and K3). Using CPCSLD, we describe the developmental changes in word syntactic categories and frequency across K1, K2, and K3. Validation analyses showed an advantage for CPCSLD over existing databases in predicting children's picture-naming performance, but not in the semantic decision task. Correlation analyses further revealed distinct developmental patterns in lexical properties: CPCSLD's syllable frequencies were highly correlated with those from other child-based databases, whereas its word frequencies showed low correlations. These findings highlight CPCSLD's sensitivity to early lexical development and its value for research on language production in 3-6-year-old children.