Aristotle’s Rhetoric defines fear as imagining, but his study of the soul, De Anima, says that imagining danger doesn’t evoke any emotions, whereas that just “thinking” about it will immediately bring on emotions.
Here is a good example of the problem with translations. In Aristotle’s Complete Works J.A. Smith translates the relevant passage as
… when we think something to be fearful or threatening, emotion is immediately produced …; but when we merely imagine we remain as unaffected as persons who are looking at a painting of some dreadful … scene. (de A. 427b21-24)
The translators of Aquinas’s commentary on De Anima also use “think” here. (Aquinas  195) The English verb “think” covers many activities, from just supposing to knowing to understanding. In Greek, however, Aristotle uses the word doxásōmen from the Greek verb dokein, which we have already seen when we talked about disgrace (ádoxía). The sense of dokein is to expect, to think to do, not just to think.
In his commentary on this passage, Aquinas notes that “mere imagining passes no judgment on things.” It’s not that imagining does not lead to feeling, but that imagining leads to making judgments, implicit or explicit, about how bad off we are or could be. We do not commit ourselves when we imagine in the same way as we do when we make judgments. Likewise, we do not commit ourselves when we just think, precisely the point Smith’s translation misses.
We fear when we judge what we see (imagine, make an image) to constitute danger. In feeling fear our body commits to that judgment. Philosophers continue to debate these issues. For example,
Judgmentalism holds that an emotional state is a combination of some cognitive component with an affect—often described as a form of pain or pleasure—and also, perhaps, with some desire. The theory thus individuates specific emotions by differences in their constitutive belief or judgment. (D’Arms and Jacobson 128)
As is typical of much modern philosophical writing, D’Arms and Jacobson jump over everyone between Aristotle and Hume. Ironically, the title of their article, “The significance of recalcitrant emotion (or, anti-quasijudgmentalism),” tells you how scholastic, as distinct from scholarly, these debates have become. One phrase in the sentence just quoted tips us off to a major source of this unnecessary complexity—”cognitive component.” For centuries much ink has been spilled over questions about how and how much pain, fear, shame, and other emotions, feeling, and even sensations constitute knowledge. What do we know when we feel pain? And how do we know? How much does knowing cause pain? The problems with such questions start with moves like inserting the intellectual act of judging between imagining and feeling pain, as Aquinas says Aristotle does.
Aristotle’s claim that we don’t feel fear when we imagine danger any more than we feel afraid when we look at a terrifying picture is questionable on both counts. Aquinas’s explication of the role that judging plays in making us afraid does not resolve these questions and alerts us to the problems that arise when we bring knowing into the complex process of hurting. Modern science and philosophy of pain are still wrestling with the same major factors in pain and fear as were Aristotle and Aquinas: sensing, imagining, and feeling. Instead of complicating the search with all the varieties “within the field of judgment … —knowledge, opinion, understanding, and their opposites” (de A. 427b25), I recommend that we keep this inquiry simple. Towards this end, as we look further into pain and disturbance, fear and shame, let’s keep In mind two insights from Aristotle and Aquinas. First, imagining is movement, “a movement resulting from the actual exercise of a power of sense.” (de A. 429a2) Second, we feel pain when we expect, when we commit to, that is, when we go “yes” or “no” to that movement.