Articles on child speech research

Research articles - Work post-PhD

My most recent presentation of the ideas in Part 1 of my thesis was at the Embodied Speech conference in Paris in 2019. My presentation was called, Speech timing phenomena: the role of child speech breathing


Building on Part 2 of my thesis, Ian Howard and I worked on how L1 speech sounds are developed using a computer model of an infant, that we called Elija:


Our most recent account of how children start to develop the inventory of L1 sounds is:


Ian Howard and I were invited to comment on an experiment reported by MacDonald et al. (2012) in Current Biology:

Research articles - 2007 PhD

If you would like to see my thesis exactly as submitted then download:

However, for more efficient printing you might prefer a single-spaced version:

Alternatively, you might like to see just the abstract and table of contents.


Part 1 of the thesis dealt with how timing relationships emerge from the breath stream dynamics of child speech rather than through being imitated (modelled). (I.e. how English, for example comes to have long and short vowels, pre-fortis clipping, 'stress-timed' rhythm, and so on.)

Part of it was summarised in:

Early versions were summarised in:


Part 2 of the thesis, dealing with the replication of speech sounds, was summarised in Messum P.R. (2007) How children learn to pronounce and in a poster Messum P.R. (2007) ASA New Orleans.

I would recommend looking at these before reading Part 2 (which is rather long), to get the gist of the argument. The argument is, of course, also developed in the 2015 Journal of Phonetics article, but the original thesis makes many points that could not be included in the article.

Teaching pronunciation

Articles that apply these ideas to the teaching of pronunciation (of English and other languages), are now all collected at the PronSci downloads site.