Introducing Phon 4

[Yvan Rose]

Dear colleagues,

In the context of the currently ongoing International Child Phonology Conference at the University of Wisconsin-Maddieson, it is with great pleasure that we are introducing version 4 of Phon. This new version is the result of three full years of focused development. At the centre of it is a brand new file format that fully integrates all the elements of the CHAT format utilized by CLAN, the main program behind CHILDES and several databases within TalkBank. Likewise, the new format of Phon also integrates the interval and point tiers required for acoustic analysis using Praat. This marks a giant leap forward in our long-term vision to maximally integrate Phon, Praat and CLAN formats and functions.

You can download Phon 4 at https://www.phon.ca. The current version is Phon 4.0.0-beta.6. While we are still in beta testing mode — we are still doing a bit of polishing— Phon 4 has been thoroughly tested among different research groups working on different languages worldwide (across the Americas, Europe and Australia). Thank you all, wonderful people!

Below we summarize the most central new and updated features of Phon 4. In addition, we are currently expanding the Consistency Analysis and will soon introduce additional measures such as the Stratified Analysis of Polysyllables, which will replace the current Word-level Analysis of Polysyllables.

We look forward to your feedback or questions. We also want to remind all of us about the importance of data sharing, the reason why we build software for the community. Publishing it through PhonBank and CHILDES amounts to a minuscule fraction of the work involved in corpus building, yet can provide the most visibility to your research. Data sharing also offers of the most tangible long-term contributions to our field of research. And we have tools and knowledge to help with every aspect of corpus creation. Make sure so be in touch!

Warm regards,
Yvan Rose & Gregory Hedlund, also on behalf of Brian MacWhinney

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Phon 4: Summary of new and updated features

1. Modern, document-style transcript editor. Phon 3 showed one utterance at a time. The new editor in Phon 4 displays a single scrollable transcript — including an editable session header (date, media, participants, languages). The whole application also has a fresh look with light and dark modes that follow your computer's appearance.

2. Native CHAT/CLAN editing. Phon 4 can open, edit, and save CHILDES .cha files without losing information. CHAT main-line conventions (terminators, overlaps, retracing, events) and dependent tiers (%mor, %gra, %wor, …) are fully supported, with on-screen code-entry grids and CLAN-style data format validation.

3. Two-way Praat / TextGrid exchange. Acoustic analysis is now built in — spectrogram, pitch, formants, and intensity are computed inside Phon, with no separate Praat installation needed. You can import Praat TextGrid annotations through a guided wizard and export Phon's timeline tiers back to TextGrid.

4. Inter-Transcriber Agreement. A new built-in reliability analysis for two-transcriber studies: Percent agreement and Cohen's Kappa, reported per word with separate consonant and vowel tables. Words missing a second transcription are listed separately rather than scored. (Developed in collaboration with Donna Thomas at the University of Sydney.)

5. Rebuilt waveform and segmentation. The waveform scrolls smoothly and stays sharp at any zoom, and segmentation and playback follow a smooth marker. Large media files load in the background, so the editor no longer freezes when opening a session transcript.

6. Faster with large corpora. Long longitudinal sessions open without stalling, and the application starts up noticeably faster.

7. Smarter pattern searching (Phonex v4, currently in late-stage development). Searches can now match across aligned IPA Target ↔ IPA Actual pairs, so phonological processes — deletion, epenthesis, metathesis, assimilation — can be found in a single query instead of two.

8. "Schwa" an AI conversational assistant built into the Session Editor (currently in early-stage development). It can answer questions about your data, run queries and analyses (e.g. PCC/PVC), and help make transcript edits. It works with cloud AI services or fully offline, for privacy-sensitive data.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -