The specification stage addresses what is tested: content standards. It is concerned with how test content relates to the qualitative descriptive scheme or 'horizontal dimension' of the framework. This includes the question of how test tasks relate to social activities in the target language use domain to which test results are intended to generalise. According to the action-oriented perspective adopted by the CoE, the linguistic resources required to understand textbooks and produce essays in the educational domain are distinct in important ways from those needed to, for example, engage with sacred texts and participate in religious services ('public domain') or to interpret recipes and cook meals ('personal domain'). Results from a test that addresses one context may not readily generalise to another. Responding to criticism that the CEFR lacks the level of detail required to build test specifications (see for example Weir 2005), North (2014) and the other CEFR authors stressed that the framework is not itself a content standard, but a generative "apparatus to develop a differentiated standard appropriate to the context" (p.62). The CEFR offers guidance for building contextually relevant standards: what [specified] learners have to learn to do in order to use a [specified] language (or languages) for [specified forms of] communication and what knowledge and skills they have to develop so as to be able to act effectively [in specified contexts]. Contrary to Kaftandjieva (2007), who believed that limitations in the coverage of the 53 illustrative scales by a test would invalidate any claim of linkage for that test, North (2014) emphasised that tests might draw on the framework selectively. Some of the elements to be specified might be found in the CEFR, but users should add others according to need. In relation to languages for academic purposes, he listed eight CEFR scales as relevant to academic language use, but suggested these could provide no more than a point of departure for specifying the content domain of academic language use. He stressed that because, "two examinations may both be "at