Drawing on nearly 20 years of study of a variety of languages in North East India, from the Tai and Tibeto-Burman families, this paper examines the issues involved in studying those languages, building on three well established principles: (a) tones are categories within a language, and the recognition of those categories is the key step in describing the tonal system; (b) in at least some languages, tones are a bundle of features, of which (relative) pitch is only one; and (c) tones may carry different levels of functional load in different languages. I will discuss the use of historical and comparative data to assist with tonal analysis, while raising the possibility that the tonal categories of individual words may vary from one language variety to the next. Different approaches to marking tones, for linguistic transcriptions, presentation of acoustic data (F0) and in practical orthographies are discussed, along with the effect of intonation and grammatical factors such as nominalisation on the realisation of tones.