Phantasia is the perceptual or noetic stimulus for the activation of desire and, therefore, for action. Nussbaum's discussion brings out this feature of Aristotle's concept of phantasia. But on one significant point I must register disagreement with her treatment, although a full discussion of the point would be out of place here. 12 Nussbaum argues that phantasia as an interpretive faculty (the most important role it plays in the psychology of action) is to be -11- contrasted with phantasia as a faculty which preserves and presents representations of the content of perception. Phantasia as a faculty of 'images' of this sort is in her view a part of a regrettably lame empiricist psychology built around the theoretical device of mental images. By contrast, she finds the interpretive kind of phantasia philosophically interesting. She therefore attempts to pry apart the promising texts where the interpretive function of phantasia is prominent from those passages where the context demands that images be a part of the concept, treating these two aspects of phantasia as alternative types. As an exegetical approach this is dubious. In the psychological treatises Aristotle speaks of the faculty of phantasia as though it were one capacity of the soul. It would be an unfortunate muddle on his part if he used one technical term to refer to two distinct and mutually exclusive psychological powers or processes. It is preferable, if at all possible, to suppose that he had a single faculty in mind throughout the psychological works, a faculty which combined both the interpretive and representational features which are required of phantasia at various places in those works. This approach would also enable us to avoid the need to suppose (as Nussbaum sometimes does) that the interpretive phantasia is a faculty which goes to work on uninterpreted data provided by aisthêsis. 13 If, like perception, phantasia supplies images which resemble their objects and also interprets them, we will find it easier to understand why Aristotle emphasizes the similarity of perception and phantasia (as he does at De Anima 431a15, 432a9-10 and in De Anima 3.3 where phantasia is a change in the soul (kinêsis) derived from aisthêsis). Such a reading of phantasia and phantasmata would require us to suppose that Aristotle believed that some interpretive images are needed for thought. For he clearly says that phantasmata are required in thought: 'and when it thinks (theôrêi) it must at the same time think some phantasma.' 14 This statement immediately precedes a statement that phantasmata are similar to aisthêmata. Aristotle is treating thought (as he often does) as a combination of basic concepts all of which are ultimately derived from perceptual experience. 15 Since he conceives of thought in this way, we would not expect him to draw a sharp contrast between the representational images of perception and the discursive or propositional interpretations of them. While the details of Aristotle's theory of phantasia -12- are problematic, it will not be helpful to begin, as Nussbaum does, with a sharp dichotomy between perceptual images and thought. Rather, we should try to understand how Aristotle combined the two in a single, although somewhat inchoate, concept of phantasia.
— Brad Inwood
Inwood's objection to Nussbaum is precise, and the precision matters beyond the textual dispute. Nussbaum wants to rescue Aristotle from an embarrassing empiricism by separating the interpretive work of phantasia — philosophically interesting, action-guiding — from its image-making work, which she treats as a theoretical liability. Inwood argues this split is exegetically indefensible: Aristotle uses one term because he means one faculty, and the interpretive and the representational are not alternative modes but co-present features of a single capacity.
What hangs on this is substantial. If you follow Nussbaum's division, you end up with a hierarchy: the image is raw material, interpretation is where the real psychology begins. Thought gets to stand a little apart from the sensory substrate, working on data rather than through it. Inwood's reading refuses that separation. Phantasia, because it is a movement of the soul derived from perception (kinêsis ek aisthêseôs), already interprets in imaging; it does not receive unprocessed data and then appraise it. The phantasma required for thought is already a shaped, resemblant, evaluatively charged thing — not neutral material awaiting cognitive upgrade. Aristotle's insistence that thought must simultaneously think some phantasma is, on this reading, not an awkward concession to embodiment but the actual architecture: thinking runs through images that already carry the world's weight, not images that wait to be told what the world means.
Brad Inwood·Ethics and Human Action in Early Stoicism·1985