With a background in sociolinguistics, this paper presents the theoretical, methodological and epistemological issues the author was forced to negotiate in conducting the apparently simple task of investigating a few dozen people's attitudes towards the English influence on their own language. The paper discusses the fundamental epistemological shortcomings of different approaches to attitude research. Three different orders of analysis are attempted and reflected upon, a standardized quantitative analysis, a discourse analysis and a deconstructive, nonessentialist analysis. It is the author's firm believe that this critical examination of methods is on the one hand essential for academic approaches to attitude research, and on the other essential in informing the public - you and me - about the mechanisms of opinion polls which underlies so much modern political work. In this respect, the papers "failure" to set up a new and improved approach to attitude research, and its content to point out the shortcomings of the current approaches, may not be a failure as much as a conscious plea to do away with the notion of a objective or neutral investigation of opinions.