APPROACHES TO LANGUAGE Prescriptive -> judgements about how speakers should use the language. Definition of the standard variety of the language
Descriptive -> a scientific description of
phenomena, without judgements about correctness or appropriateness THE DESCRIPTIVE APPROACH Based on the observation of data and
The formulation of a hypothesis
The data are used to confirm or reject the hypothesis SOURCES OF DATA The data we can be obtained from different sources: Our own intuition. We can create data using our knowldege of a language. Interviews and questionnaires. We can create data asking questions to speakers of the language. A corpus. This is a collection of texts. They represent natural occurring data. PROPERTIES OF A CORPUS A corpus can be seen as a sample of the language (because its impossible to study the whole of the language). Like any sample, a corpus must be representative. The more representative it is, the more reliable your conclusions will be. REPRESENTATIVITY What makes a corpus more representative? The number of words. The more words it has, the more representative. The number of text sources. The more sources your corpus consists of, the more representative it is. It is better to have 100 words from 100 sources than 10,000 from one single source. The nature of the sources. A corpus will be more representative if it contains different text types, channels, topics, etc.