Language policy is all about choices. If you are bilingual or plurilingual, you have to choose which language to use. Even if you speak only one language, you have choices of dialects and styles. To understand the nature of this process, one needs an ecological model (Haugen 1987: 27) that will correlate social structures and situations with linguistic repertoires. Any speaker or writer is continually selecting features – sounds or spellings, lexical items, or grammatical patterns – which are significant markers of languages, dialects, styles, or other varieties of language, and which, bundled together, constitute recognized and labeled languages, like Navajo or English or Chinese, or more precisely, varieties of language like American English, or Midwestern English, or Cockney, or Indian English (Blommaert 2007), or what Blommaert (2008) refers to as “speech resources.” One fundamental fact about named varieties is that they are socially or politically rather than linguistically determined. A dialect becomes a language when it is recognized as such: recently, the prime ministers of Romania and Moldova are reported to have argued (the former speaking in French and the latter in Russian) at an international congress over whether their two varieties were one language or two. The various Scandinavian languages are close enough for speakers of Norwegian to understand someone speaking Danish or Swedish (Delsing 2007; Doetjes 2007) but they consider them separate languages; the speakers of Chinese varieties cannot understand each other, but they all agree they speak Chinese.