As seen from the interior, the entrance is a small īwān flanked by a pair of tiny chambers, both opening into Iwān and court. This group of three leads to the form evolved in the Madrasā aṣ-Ṣāḥībiyyā. At the corners are two closed chambers whose doors lead into the destroyed side naves of the building.
Mainly because a pilaster, A on the plan, is preserved at the northeast corner of the court, Sauvaget restores the missing parts as replicas of the prayer hall, which occupies the entire breadth of the south wing, thus arriving at identical elevations for the four sides of the court. The argument can be turned both ways: the preservation of one pilaster permits the restoration of the three others, or the preservation of only one indicates that there never were four. I cannot see how to classify Sauvagetʾs plan. A dār al-ḥadīt is a small madrasā; pupils and teachers live in it. In this one, it is known that they received guests like ibn Ǧubayr. It is not enough to say Orientals need nothing but a bedroll to spread somewhere. The small living rooms are what make a plan a madrasā; the open īwāns are equally essential. In Sauvaget's plan there are only two such chambers —those that are preserved— and no īwān, whereas there is an unwarranted number of common rooms.
Furthermore, symmetry may be assumed for such a small building, but only in relation to one axis, north-south, not to two axes. Most of the Syrian madrasās, all much larger, are not strictly symmetrical. There the function dominates the plans: never is a necessary part left out, nor is an existing part superfluous. The parts are arranged in a balanced correspondence, the main ones (prayer hall and īwān) indicating the axes, but not in symmetrical repetition. That is Iranian and does not appear in Syria before the third and last phase of the ʾAyyūbīd period, 634-58 (1236-1260 CE). It is achieved by the addition of something inessential, and repeating the prayer hall on both sides of the court in the dār al-ḥadīṯ would be just that.
