The Assyrian method of engraving stamps for impressing clay was practised by the old Roman potters, who marked their manufactures with the names of the owners or with the contents of the vessel. The potters clearly understood the [p036] value of movable types. On some of their lamps of clay, the inscriptions were made by impressing, consecutively, the type of each letter. These types must have been movable, and, in appearance, somewhat like the punches or the model letters of type-founders. There were some men in ancient Rome who had a clear perception of the ease with which engraved letters could be combined. Cicero, in an argument against the hypothesis of logical results from illogical causes, has intimated that it would be absurd to look for an intelligible sentence from a careless mixing up of the engraved letters of the alphabet.[5] The phrase by which he describes the assembled letters, _formæ literarum_, was used by the early printers to describe types. His argument implies, conversely, that if proper care were exercised, it would be easy to arrange the letters in readable sentences. But the speculation of Cicero did not go beyond the idea of combination. It does not appear that he thought that the letters could be used for printing. Quintilian had speculations about engraved letters. He recommended to teachers the use of a thin stencil plate of wood, on which should be cut the letters that a boy might be required to copy when learning to write. The boy who traced the characters with his writing implement would have his hand guided and formed by the outlines of the perforated letters. The curt manner in which stencil plates are noticed should lead us to think that they were then in common use. We can see that stencils of this nature could have been used, at least as an aid, in the mechanical manufacture of books; but it is not probable that they were so used. [p037]