Willis, David. 2013. 'Negation in the history of the Brythonic Celtic languages', David Willis, Christopher Lucas & Anne Breitbarth (éds.), The History of Negation in the Languages of Europe and the Mediterranean, 239-298.
Introduction about Breton, p.240-41: "Breton has also undergone Jespersen’s cycle. It has replaced an inherited stage I pattern with preverbal ni(t), illustrated for Old Breton in (8), by stage II (Neg–V–Neg) bipartite negation ne . . . ket, illustrated for Present-day Breton in (9). (8) Ni gus coucant. neg know.pres.impers certain ‘It is not known exactly.’ (Fleuriot 1964a: 268) (Old Breton) (9) N’ ouzhon ket piv out. neg know.pres.1sg neg who be.pres.2sg ‘I don’t know who you are.’ (Present-day Breton) The bipartite pattern is already dominant, although not compulsory, by the time of the earliest Middle Breton texts (1450), and becomes compulsory over the next two centuries. More recently, many dialects have begun to omit the preverbal marker ne, leading to a stage III negative pattern with postverbal ket alone. This chapter will begin by tracing these developments in some detail. It will then move on to examine the main developments in the system of indefinites, paying particular attention to indefinites found in negative and non-assertive (negative polarity) contexts. Here, a number of items have become increasingly negative in the historical period, leading to the cyclic creation of new series of pronouns for nonnegative contexts. Finally, various other contexts for negation will be considered, for instance, negative non-finite clauses and negative imperatives, all contexts where there has been considerable change over the history of the languages."
Les remerciements dans l'article suggèrent que les données de breton moderne viennent de Janig Stephens (locutrice native, trégorrois), mais l'orthographe en (9) est surprenante.