Paidologos Project


Jan Edwards
Hearing and Speech Sciences
University of Maryland

website
Mary Beckman
Linguistics
Ohio State

website

Citation Information

Edwards, J. & Beckman, M. E. 2008. Methodological question in studying consonant acquisition. Clinical Linguistics and Phonetics 22(12):939-958.

Edwards, J. & Beckman, M. E. 2008. Some cross-linguistic evidence for modulation of implicational universals by language-specific frequency effects in the acquisition of consonant phonemes. Language Learning and Development 4(2):122-156.

In accordance with CHILDES rules, any use of data from this corpus must be accompanied by at least one of the above references.

General Information

The paidologos project is a large, cross-linguistic investigation of phonological development. The project was inspired by the fact that words in different languages are made up of different sounds and sound sequences, and sounds that are common in some languages are rare in others. The focus of the paidologos project is to examine how these differences affect how children learn to speak different languages.

The name “paidologos” is a new word that we made up from Greek roots meaning “child” and “word,” in order to capture our idea of looking at children’s words in parallel across different languages. To do that, we traveled to day care centers and children’s homes in Hong Kong, eastern Japan, northern Greece, the far northeast of China, South Korea, and central Ohio to record the speech of two- to five-year-old children learning to speak either Cantonese, Japanese, Greek, Mandarin Chinese, Korean, or English. The corpora that are made available here include child and adult speech of Cantonese, Japanese, Greek, and English.

Participant Information

For each language, there are 10 males and 10 females in each age group (2, 3, 4, and 5-year olds, plus 20 adults). The children all come from middle socioeconomic status families. The adults all self-reported normal speech and hearing. All children were tested with a hearing screening (pure tone screening at 25 dB HL for 500, 1000, 2000, and 4000 Hz or otoacoustic emissions at 2000, 3000, 4000, and 5000 Hz), and norm-referenced measures of expressive vocabulary (Williams 1997), receptive vocabulary (Brownell 2000), and articulatory accuracy (Goldman & Fristoe 2000). Any child who did not pass the hearing screening in at least one ear or who scored more than one standard deviation below the mean on the norm-referenced measures was excluded from the study.

Recording & Transcription Procedure

For each language, the recorded stimuli used as elicitation prompts were spoken by an adult female native speaker, in a child-directed manner, and digitally recorded and presented at a sampling rate of 22,500 Hz.

Each trial consisted of a picture and the associated sound file, which were presented simultaneously to the participant over a laptop with a 14-inch screen. The computer program included an on-screen VU meter to help the children monitor their volume and a picture of an animal (duck or frog or koala bear) walking up a ladder on the left side of the screen to provide visual feedback to the children about how close they were to completing the task. The children were instructed to repeat each word exactly as they heard it.

The first audible response to each prompt was transcribed in a Praat TextGrid. A native speaker who is also a trained phonetician listened to the response and examined the acoustic waveform for each repetition for transcription. A second native-speaking, trained phonetician blindly re-transcribed 20 percent of the data for each language.

The children's recordings took place in a quiet room at one or more preschools in each of the four countries. The adult speakers of English were recorded in a sound booth at the Ohio State University using the same protocol and equipment. The adult speakers of Japanese were recorded in a quiet room at Daito Bunka University, using the same protocol and recording equipment. The adult speakers of Greek were recorded in a quiet room at their home, or in the home of the interviewer, using the same protocol and recording equipment. The adult speakers of Cantonese were recorded in a quiet room in the Chinese University of Hong Kong using the same protocol and recording equipment.

Naming Conventions

Each file is meaningfully named with a sequence that indicates language, age, sex, list, and participant ID, as follows:
1) – The initial character indexes the language, with "c" for Cantonese, "e" for English, "g" for Greek, and "j" for Japanese.
2:3) – The next two characters give the child's age, within a 6-month band. The number represents years, while the letters "a" and "b" indicate whether they are in the first or second half of that year, respectively. That is, for example, the child whose productions are in files c2at01fw_canw211 and c2at01fw_canw212 is a "young" 2-year-old, aged between 2;0 and 2;5, whereas the child whose productions are in files e5bt25fw_enrw111 and e5bt25fw_enrw112 is an "old" 5-year-old, aged between 5;6 and 5;12. While finer divisions of the children's ages than these 6-month bands are not permitted, it may be useful to know that the two digits after the "t" are the identification numbers for the 20 or 21 children within each year band, and these numbers are ordered by age, so that c2at01fw is the youngest Cantonese-speaking two-year-old and e5bt25fw is the oldest English-speaking five-year-old. In the case of the adults, a "g" will replace the “a” or “b”.
4) – The fourth character – “t” – stands for typically developing, and is constant across these files because there are no developmentally atypical participants.
5:6) – The following two-digit identification number indicates the individual speaker in each age group, in order from youngest to oldest.
7) – The seventh character, either "f" or "m", after the child's identification number encode the child's gender.
8) – The underscore signifies that everything following it indicates language, session, and list.
9:10) – These characters are a two-letter language code (ca- Cantonese, gr- Greek, en- English, ja- Japanese).
11:12) These stand for either “nw” – nonword or “rw” – real word (only the real words are included in these corpora).
13:14) This is the list number.
15) This is the block number (1 or 2).

There are two sets of files for each child, for the two lists of words that were presented for repetition. The word lists were designed to elicit the three words starting with each target consonant-vowel sequence in such a way that the different tokens that were stimuli were distributed across lists and children, so that no child was presented with all three words in the same block and each of the stimulus words for a sequence was taken from a different position in the series that the experimenter had recorded.

Acknowledgements

This project was supported by National Institute for Deafness and Other Communication Disorders (NIDCD) funding between 2003 and 2010.