The essays in this work focus on problems with logic-based approaches to human reasoning and show how a probablistic approach can provide a more psychologically plausible, computationally viable and philosophically respectable account of human inference. The author's probablistic approach shows that although people's behaviour appears irrational from a logical point of view, it is entirely rational from a probablistic perspective. There are new chapters on the emerging empirical support for probablistic models and on the cognitive mechanisms, such as neural networks, that may implement them in the mind. Moreover, the presentation of optimal data selection is simplified reflecting the author's experiences with teaching this material at the advanced undergraduate level.