Exploration and Inference

Around seemingly self-protecting methods of statistical inference, various ‘external uncertainties’ are lurking. They may reduce the practice of inferential techniques to the status of an exploratory activity. Conversely, exploration, consciously practiced, can be a liberated form of inference, partly dispensing with the strict requirements of the latter and renouncing its built-in protection — as far as these are undermined by external uncertainty. Such interchanges between Inference and Exploration show the important role of the framework of induction as an ‘environment’ surrounding the statistical act of inference and supplying it with supplementary information.