‘But’ how do we reason with it: An experimental investigation of the implicature stemming from ‘but’

• We investigated ‘p but q’ sentences as distancing-contrastive connections.
• We found that people generally make the inference induced by ‘but’.
• The expected ‘nevertheless’-conclusions are harder to make than ‘so’-conclusions.
• The content of the arguments plays an important role in ‘p but q’-constructions.
• People consider the conclusion from the p-argument spontaneously.

In two experiments, we aimed to investigate whether people truly make the inference induced by ‘but’ – combined with ‘so’ and ‘nevertheless’ – in ‘p but q’ sentences constructed as distancing-contrastive connections. In Experiment 1, our participants were presented with ‘p but q’ sentences that contained both sensible and absurd arguments and were instructed to indicate the appropriate ‘so’- or ‘nevertheless’-conclusion. We found that, while people do grasp the pragmatic meaning of ‘but’, the content of the arguments plays a very important role: when a sensible and an absurd argument were combined, the majority of participants based their answer on the sensible argument. We also found that the expected ‘nevertheless’-conclusions from p are very hard to make since they require participants to overrule the inference stemming from ‘but’. In Experiment 2 we further explored the role of the content by explicitly asking participants to explain/give reasons for their answers. Consistent with our expectations, we found, that whenever participants did not infer the appropriate conclusion from ‘but’, they referred to the content of the arguments. This means that people spontaneously consider the implication from the p-argument as well.