Hidden interlocutor misidentification in practical Turing Tests

Based on insufficient evidence, and inadequate research, Floridi and his students report inaccuracies and draw false conclusions in their Minds and Machines evaluation, which this paper aims to clarify. Acting as invited judges, Floridi et al. participated in nine, of the ninety-six, Turing tests staged in the finals of the 18th Loebner Prize for Artificial Intelligence in October 2008. From the transcripts it appears that they used power over solidarity as an interrogation technique. As a result, they were fooled on several occasions into believing that a machine was a human and that a human was a machine. Worse still, they did not realise their mistake. This resulted in a combined correct identification rate of less than 56%. In their paper they assumed that they had made correct identifications when they in fact had been incorrect.

Linguistic experiments and ordinary language philosophy

J.L. Austin is regarded as having an especially acute ear for fine distinctions of meaning overlooked by other philosophers. Austin employs an informal experimental approach to gathering evidence in support of these fine distinctions in meaning, an approach that has become a standard technique for investigating meaning in both philosophy and linguistics. In this paper, we subject Austin's methods to formal experimental investigation. His methods produce mixed results: We find support for his most famous distinction, drawn on the basis of his `donkey stories', that `mistake' and `accident' apply to different cases, but not for some of his other attempts to distinguish the meaning of philosophically significant terms (such as `intentionally' and `deliberately'). We critically examine the methodology of informal experiments employed in ordinary language philosophy and much of contemporary philosophy of language and linguistics, and discuss the role that experimenter bias can play in influencing judgments about informal and formal linguistic experiments.

Economic crises in England, 1270-1520: a statistical approach

The chapter examines how far medieval economic crises can be identified by analysing the residuals from a simultaneous equation model of the medieval English economy. High inflation, falls in gross domestic product and large intermittent changes in wage rates are all considered as potential indicators of crisis. Potential causal factors include bad harvests, wars and political instability. The chapter suggests that crises arose when a combination of different problems overwhelmed the capacity of government to address them. It may therefore be a mistake to look for a single cause of any crisis. The coincidence of separate problems is a more plausible explanation of many crises.

Science for Humanitarian Emergencies and Resilience (SHEAR) scoping study: Annex 3 - Early warning system and risk assessment case studies

This report provides case studies of Early Warning Systems (EWSs) and risk assessments encompassing three main hazard types: drought; flood and cyclone. The case studies are taken from ten countries across three continents (focusing on Africa, South Asia and the Caribbean). The case studies have been developed to assist the UK Department for International Development (DFID) to prioritise areas for Early Warning System (EWS) related research under their ‘Science for Humanitarian Emergencies and Resilience’ (SHEAR) programme. The aim of these case studies is to ensure that DFID SHEAR research is informed by the views of Non-Governmental Organisations (NGOs) and communities engaged with Early Warning Systems and risk assessments (including community-based Early Warning Systems). The case studies highlight a number of challenges facing Early Warning Systems (EWSs). These challenges relate to financing; integration; responsibilities; community interpretation; politics; dissemination; accuracy; capacity and focus. The case studies summarise a number of priority areas for EWS related research: • Priority 1: Contextualising and localising early warning information • Priority 2: Climate proofing current EWSs • Priority 3: How best to sustain effective EWSs between hazard events? • Priority 4: Optimising the dissemination of risk and warning information • Priority 5: Governance and financing of EWSs • Priority 6: How to support EWSs under challenging circumstances • Priority 7: Improving EWSs through monitoring and evaluating the impact and effectiveness of those systems