The assumption that there are two types of models is put forward and tested experimentally. A positive result of checking the given object with its "hard" model is usually achieved when the object is identical with the model, while checked with its "soft" model the result is positive provided the object is identical with or close to the model. The closeness allows to put forward a hypothesis concerning the object character to be further checked with its "hard" model. The identifying does not determine the closeness of the object to the model through considering their co-incident features. Another principle is used: the model and the object (rather its perceptive reflection) are compared as integral, inseparable into parts, units. The extent of their closeness determines the check result and finally brings about a pertinent hypothesis. The problem suggested is teaching schoolchildren to use both "soft" and "hard" models.