A dominant way of learning and proving concepts like the Pythagorean theorem in schools is through visual illustrations and visually organized assignments. Visually impaired students (VIS) cannot see and thus cannot benefit from any such illustrations alone. Often, objects are then used to help the VIS achieve a tactile and haptic perception of the object, that represents the mathematical concept. However, there are so many important details embedded in transforming a visually available object to be of specific scientific relevance for tactile and haptic exploration. If one does not pay attention to such details, the object becomes at best useless. We have, in a video ethnographic project, followed VIS in their everyday lives in Danish schools, focusing on objects as situated resources. A recurrent phenomenon is when VIS receives help from assistants in solving assignments and the assistant guides the hands of the VIS towards an object as a representation of the mathematical concept. I focus in this paper on ways the touched object represents the Pythagorean theorem and I show how the assistant guides the hand of the VIS to the object but that the VIS eventually is unable to understand the concept. I show and argue that the failure cause is the lack of detailed haptic exploration of the triangles full shape and details of its angles. To support the argument, I draw on Gurwitsch theory of the relation between ‘gestalt contextures’ and ‘functional significance’ in the praxeological respecifications that Garfinkel advanced. This research contributes to ethnomethodological studies of science by critically examining the use of objects for VIS in math classes and by showing in practical details how the ethnomethodological version of gestalt theory is a useful framework for understanding object-centred sequences.