Keynesian Uncertainty and the Weight of Arguments