Data have been obtained to support our hypothesis that "mild", not strict standards in cognition when they are whole (not made up of elements), help the hypothesis about the nature of the object being perceived. A technique of displaying on a computer monitor an unfocused object which later became still more unfocused, was used to study microgenesis of cognition. It has been established that when the figure being displayed is close to the "mild" and whole standard, subjects put forward progressively improving hypotheses and eventually solve the cognitive problem. If the figure, however, does not approach the "mild" standard, the above mentioned hypotheses are not put forward and the task is not solved as a rule. It has been experimentally proved that these failures cannot be explained by the difference between separate elements of the figure and elements of the standard if the standard is regarded as made up of elements, not whole.