S-images in the sequence are connected to other s-images before and after it with transformations. Transformations, like functions, take arguments to specify their behavior. Making topological changes of this kind to imagined physical systems has been shown in earlier work to be useful in problem solving. (Griffith et al., 2000; 1996). Table 1 shows some examples of transformations.
| Table 2 | |||||
| Transformation name | arguments | ||||
| move-to-location | object, new-location | ||||
| move-to-touch | object, object2, new-location | ||||
| move-above | object, object2 | ||||
| move-to-right-of | object, object2 | ||||
| move-below | object, object2 | ||||
| move-to-left-of | object, object2 | ||||
| move-in-front-of | object, object2 | ||||
| move-off-s-image | object, location | ||||
| move-to-set | object, object2 | ||||
| rotate | object, direction | ||||
| start-rotating | object, direction | ||||
| stop-rotating | object | ||||
| start-translation | object, direction | ||||
| stop-translation | object | ||||
| set-size | object, new-size | ||||
| add-element | object, location (optional) | ||||
| remove-element | object | ||||
| decompose | object, number-of-resultants, type | ||||
| scale | object, new-size | ||||
These transformations control normal graphics transformations such as translation (move-to-location, move-to-touch, move-above, move-to-right-of, move-to-left-of, move-below), rotation (rotate), and scaling (set-size). In addition there are transformations for adding and removing elements from the s-image (add-element, remove-element). Certain transformations (start-rotating, stop-rotating, start-translation, stop-translation) are changes to the dynamic behavior of the system under simulation. For example, rotate changes the initial orientation of an element. In contrast start-rotating sets an element in motion. In simulation, if something is touching it, there would be friction.
Glasgow and Papadias (1998) also came up with primitive visual functions for spatial reasoning. They are: retrieve, put, find, delete, move, turn, focus, unfocus, store, and adjacent (p. 186). My choice of transformations will be constrained by the findings of this work and others like it (Thagard & Hardy, 1992 added the primitive ``surround'' to this ontology.)