Frequency Counts

How often are particular words used in speech to and from children? Type comma-separated words, pick a collection, and compare usage frequency (parts per million tokens) across ages, speakers, corpora, and children. Counts are over all utterances attributed to a speaker, matched case-insensitively.

Loading data — the first load can take a few seconds…

Points show word frequency per age bin (occurrences / total tokens for that speaker role, × 10⁶); curves are token-weighted LOESS smooths that include ages where a word did not occur (frequency 0). Bins with fewer than 100 tokens are hidden. When specific children are selected, panels show each child separately.

Note

Counts come from the childes-db 2026.1 type-frequency table (one row per word type per speaker per transcript); denominators are total tokens per speaker per transcript from get_speaker_statistics, aggregated over the same selection. Words are matched on the NFC-normalized, lowercased gloss. The first query initializes an in-browser database (~10 s); later queries fetch only the ~2 MB shard containing each word.