A model [is] something other than an induction. For induction is a method [that proceeds] from particulars to what is universal. But a model produces what is similar, and it is more knowable than what is similar and it is less [knowable] than the knowable. This is why induction does not arise through one thing. But it suffices for a model to be through one thing. And while induction is more appropriate for the dialectician, a model [is more appropriate] for an orator.[1] However, the dialectician too has made use of models for the sake of the clearness of what is being said, if it were not well-known; so as to find out about it through the similarity. For to ask like this: "is the knowledge of the contraries the same, as the perception [of contraries is the same]?" is paradigmatic. There are differences between paradigm and parable, since the paradigm seems to provide something already produced to show something that had not been produced yet and, for this reason, is unclear. By contrast, a parable is an image indicating a similarity related to something that has been produced. [2]
Note that[3] models are not admitted in everything and [that they do not] fit well into what is proposed. For [in that case] they would not be examples but identity. However, we must pick out what is advantageous and for what purpose it is used, letting the rest go. The Patriarch [sc. Moses], for example, says about the Savior, "reclining [...] like a lion";[4] we do not apply everything to Christ, but when receiving what is unconquerable, fearful, and kingly, even though there is something else, we ignore [literally: ride past] what is beastliness. And when [he] says "I shall meet them like a she-bear bereft [sc. of her cubs]",[5] we only shall accept what is revengeful and not some other aspect of what is characteristic of a bear.