Quote From phdbug:
What would sample size be for an ethnographer who spends 10 years studying kinship patterns in one community? Or, what would sample size mean for a critical discourse analyst who is studying two texts, or for a semiotic theorist who is studying early childhood literacies of two children?
Just wanted to give my thoughts. In the cases you've outlined the sample size issue can be thought of in terms of analysis, and in qualitative data analysis (regardless of the exact method used) the important thing is to keep analysing until data saturation is complete, that is, until no new issues are arising from the data.
So, in the example of the ethnography in one community, you have to make sure you've pulled everything from the data you've obtained.
In the discourse issue - you need to make sure that the entire text has been extensively analysed, not just sections of it.
Does that make sense?