Phantasia and Perceptual Realism in Aristotle

Perceptual Unreliability and Phantasia in Metaphysics 4.5

Perceptual errors are commonplace. We know we mistakenly perceive a straight stick in the water as bent or a large tower in the distance as small when we know what they really look like. But what if their “true appearances” are dubious as well? Sometimes we do confuse true and false appearances, but if a thing’s incompatible appearances were deemed epistemically equivalent, there would be no “true perception” at all. Perception would then not be an epistemically reliable approach to reality: we can never infer “it is red” from “it looks red”.

Since antiquity, when philosophers have assessed perception epistemologically, illusory experiences have invoked serious skepticism about the epistemic reliability of perception. For instance, the existence of perceptual error led Plato to draw the line between the constantly changing physical world and the stable ideal world [16]6, Epicureans to divide infallible sense impressions from fallible beliefs [17]7, and Stoics to set apart cognitive and non-cognitive impressions [17]8. Additionally, perceptual unreliability has sometimes incurred an outright rejection of the reality of sensible objects, i.e. perceptual anti-realism, epitomized by

4  Rapp 2001 and Frede 2020 propose two-concept interpretations of phanta- sia in response to this tension, suggesting their failure to unify these two senses. My account of non-veridical phantasia will unify these two senses, though I admit that this cannot be its only sense.

5  Caston 1996 maintains a similar view on the purpose of phantasia, but along a different line. He claims that dealing with perceptual error is the purpose of the whole chapter De Anima 3.3, while for me this is only the purpose of the aisthesis-phantasia distinction. Phantasia’s difference from doxa (427b16-24; 428a18-24), I suggest, serves to indicate that phantasia as a sensory faculty is without assent or commitment characteristic of rational faculties.

6  Phaedo 78b4-81a2, Republic 509d1-511e3 in Cooper 1997.

7  E.g., Sextus Empiricus, Against the professors, 8.63 in Long & Sedley 1987.

8  E.g., Diogenes Laertius 7.46 in Long & Sedley 1987.

the Heraclitean flux doctrine and the Protagorean thesis that “all appearances are true” [16]9. Either assuming that sensible objects are in a constant state of flux or that contradictory sense impressions are epistemically equivalent, the universe of sensible objects cannot be anything more than a fantasy or fabrication. Perceptual anti-realism agrees with the unreliability of perception because the lack of reality explains why sensible objects produce contradictory perceptions and so cannot provide the basis for epistemically reliable information.

The issue of perceptual unreliability confronts Aristotle as well. Aristotle acknowledges that perception is sometimes deceptive (Metaphysics 1010b19-30, De Anima 428a12- 15, 428b2-5), but decides to explain this phenomenon differently. Notably, Aristotle refuses to equate perception and phantasia, as the Epicureans and Stoics do, or conflate perception, phantasia, and doxa, as Plato does10. Nor does he compromise the reliability of perception. Rather, he holds that perceptual reliability can be saved if we manage to distinguish reliable perception from the unreliable type of sensory experiences – phantasiai/phantasmata.

In his criticism of the Protagorean thesis that “all appearances are true” in Metaphysics 4.5, Aristotle blames this absurdity on the confusion between perception and phantasia (1010b1-14): As for truth, to show that not every phenomenon (phainomenon) is true: first, even if perception, at least of what is proper, is not false, still phantasia is not the same thing as perception. Next, one may legitimately be surprised that they [Protagoreans and Heracliteans] should find perplexing the question whether magnitudes and colors are such as they appear to those who are at a distance or those who are near, and by the healthy or the sick; or whether what appears to the weak or the strong is heavier; or whether what appears to the sleeping or the waking is true. For it is obvious that they do not really consider it so: at any rate if someone in Libya supposes himself one night in Athens, he does not set off for the Odeon. Again, as for the future, as Plato also says, the opinions of a doctor and an ignorant man are surely not equally authoritative, as for instance on the question whether someone is or is not going to be healthy. (1010b1-14) [18, 19]11.

If all experiences were true, there would be no objective criteria of truth with which to adjudicate contradictory experiences, and all experiences would be true for different

9  Theaetetus 152c1-e10 in Cooper 1997.

10  Cf. 428a5-16 and notes7-8; cf. 428a24-b9 and Plato’s view at Sophist 264a6.

11  Kirwan 1993’s translations and modified according to Ross 1924.

people and relative to different perspectives. By erasing the difference among all experiences, the Protagorean thesis makes the reliable type of perception, i.e. proper sensation, identical with fallible experiences which are phantasiai. However, even if proper sensations are almost always true, phantasia must be of a different type for Aristotle. He evidences the fallibility of phantasia by the fact that those who possess phantasiai are non-committal to their experiences: imagining he is in Athens, the Libyan is conscious of its imaginary nature and does not really believe it is the case, just as people who can discriminate scientific from unreliable medical opinions (1010b13-14). In their practical dealings, those Protagoreans and Heracliteans presuppose the unreliability of some experiences and are alert to the difference between reliable and unreliable experiences; it is only theoretically they treat all experiences indifferently.

Therefore, we see that Aristotle has no fundamental doubt about the reliability of perception – there must be some basic kinds of sensory experiences mirroring things as they are. What he has to do is harmonize the reliability of perception with its capricious character, and he means to do it with phantasia, which he takes to be derivative from perception. In Metaphysics 4.5, however, Aristotle does not expound on the nature of phantasia and its specific difference from perception. It is only in De Anima 3.3 that Aristotle embarks on this task, where he details the aisthesis- phantasia distinction (428a5-16) and draws the definition of phantasia from its causal dependence on perception (428b10-429a9). In what follows, I revisit these texts and demonstrate my view that the distinction in question stands between veridical perceptions and fallible experiences, based on Aristotle’s deep-seated belief in perceptual realism.

The Aisthesis-Phantasia Distinction in De Anima 3.3

The immediate reason Aristotle distinguishes phantasia from perception in the context of De Anima 3.3 is to ascertain what it is to be phantasia (ti esti). Although phantasia is too closely associated with perception to differ from it (428b10- 14), they should not be identical if phantasia is to be anything at all. Moreover, it has to be defined independently and with reference to its own objects (antikeimena) (Independent Objects Principle, IOP), as Aristotle stipulates regarding psychic faculties (415a14-20), if it is to be an individual psychic faculty at all. In other words, his argument for the aisthesis-phantasia distinction will hold true only if it can somehow tell the difference between their corresponding objects.

This task, however, is complicated by Aristotle’s idea that phantasia and perception are physiologically inseparable, which can be evidenced by at least two facts. First, Aristotle defines phantasia according to its causal relation to perception since phantasia is “of those things of which perception is” (428b12-13) and the act of phantasia, as the psycho-physical motion resultant from the act of perception, is similar to the latter (428b14). Perception and phantasia have similar and causally-related activities because they share the same causal antecedents – aistheta, which are the external agents or ‘movers’ of perceptual activities (aisthesis) and contents (aisthemata); they are similar because the original and residual motions caused by the same agents and occurring in the same place (central sense organ) are similar (De Somno 456a1-7)12. Phantasia has phantasmata (fantastic experiences or their material bearers) as its contents (428a1- 2) [15, 20, 21, 22, 23]13, whereas phantasmata are not objects in the sense of external agents, because phantasiai, as said, share the same external agents or causal antecedents with perception (428b12-13). That is why commentators rightly state that phantasia does not qualify as an individual faculty but a quasi-faculty: it cannot be defined by appeal to a special kind of external object as its causal antecedent [11, 24]14.

Second, Aristotle elsewhere concedes that phantastikon and aisthetikon are the same faculty with different “beings” or essences (De Insomniis 459a16-17), which seems to stem from the difficulty in dividing phantastikon from other parts of the soul at De Anima 3.9, 432a31-b3 [25]15. This identification can be explained by the fact that perception and phantasia share the same physiological apparatus – the same set of sense organs and causal antecedents – though performing different functions. Their functional difference somehow squares with their physiological inseparability.

In this situation, the only way to assign different “objects” for perception and phantasia is to identify different experiential contents, namely, to develop a justifiable account of the aisthema-phantasma distinction. That is, to modify IOP

12  Aristotle claims even proper sensation is accompanied by self-conscious- ness in the central organ (De Anima 425b12-25; De Somno 455a12-b2), indicat- ing that for any perception to occur, motions in peripheral sense organs, i.e. ais- themata, must persist and lead to motions in the central organ. Phantasmata, in their material aspect, are said to be lingering motions in peripheral sense organs, which are carried back into the central organ (De Insommnis 460b2-3, 460b28- 461a8).

13  Caston 2021 appropriately distinguishes the physiological movements themselves (phantasmata) and their imagistic visualization to emphasize the physiological basis of phantastic experiences.

14  For the ineligibility of phantasia as a separate faculty see also Frede 1992, 281 and Turnbull 1994, 322.

15  “And there is the phantastikon, which differs from them [nutritive and per- ceptual faculties] all in being, though there is considerable difficulty in saying–if one is going to posit separate parts of the soul–with which of the others it will be the same or from which of the others will it differ.” (432a31-b3, translated by Shields 2016 and modified.) (Modified Independent Objects Principle, MIOP). This scheme is evidenced by the preliminary definition of phantasia at 428a1-2: phantasia enables us to entertain “a certain phantasma”. I hold that Aristotle establishes this when he describes the difference between veridical perceptions and non-veridical experiences, and this, more than anything else, is their real difference. In what follows, I will show that among Aristotle’s arguments for the aisthesis-phantasia distinction (428a5-16), the fourth one (428a12-15) conveys this difference, with the other four being, strictly speaking, insufficient. For the sake of reasoning, let me address these arguments in order.

That it [phantasia] is not perception, then, is clear from the following. [A1] For perception is either a potentiality or an actuality, such as sight and seeing; yet something may also appear even when neither of these is present, as in sleep. (428a5-8) [25]16 [A2] And there is also what we said earlier, that visual images (horamata) appear even to those whose eyes are closed. (428a15-16) Aristotle taps into dreams in [A1] to signal that there are cases where an act of phantasia is not identifiable as or reducible to an act of perception. When we ‘see’ in a dream, we are not using our eyes, since bodily affections in sleep incapacitate the primary sense organ and we temporarily lose the capacity and actuality of sight (De Somno 455a8-10). Thus, ‘seeing’ in a dream does not count as seeing by sight. Still, we experience what we ‘see’ by virtue of phantasia. Extending the sense of phantasia to visualization broadly speaking (including memory and dream), [A2] offers a cross- reference to [A1] to express the same point. That is, when we are dreaming, memorizing, imagining, or hallucinating in a faint (De Somno 456a11-12), etc., we have visual experiences without using our sight by virtue of phantasia. In general, since sensory experiences occur without sensible objects as well, and phantasia is responsible for these experiences, phantasiai (or phantasmata) are different from perceptions, experiences concurrent with external objects.

Although [A1-A2], by pointing out cases of one without the other, is sufficient for claiming that perception and phantasia are not identical, it is still insufficient for proving their exclusive distinction as demanded by MIOP. For it does not prevent their overlap. This line of reasoning requires phantasia to be defined as visualization or imaginative experience free of external objects, but phantasiai are arguably more than imaginative experiences. At 428a12-15 and 428b2-5, Aristotle copes with cases like seeing a man indistinctly and mistaking the size of the sun, suggesting that he does not restrict the notion of phantasia to ‘visualization’

16  Translated by Shields 2016 and modified.

but holds it responsible for sensory experiences both with and without external agency. He is ready to explain phenomena such as illusions and indistinct perceptions with phantasia by analyzing the ordinary uses of ‘phainesthai’ and ‘phantazesthai’ in our opaque, distorted, and misleading perceptions [2, 14, 25, 26]17. Since phantasia is more than visualization, [A1] and [A2] are inadequate for the aisthesis- phantasia distinction. [B] Further, perception is always present, though phantasia is not. If these were the same in actuality, it would be possible for phantasia to belong to all beasts; but this does not seem to be the case. For instance, it belongs to the ant or the bee, but not to the grub. (428a8-11) [B] claims that perception belongs to all animals (“always present”) while phantasia is not thus ubiquitous, so they cannot be the same faculty. Otherwise, all sentient animals would necessarily possess phantasia, given that animals are sentient by definition. Aristotle believes that not all animals have phantasia: there are sentient animals with no phantasia, like grubs. While [A1] and [A2] exploit examples of ‘pure’ phantasia in which we find no sense- perceiving, like dreams, [B] turns to those ‘pure’ perceptions, like grub’s sensation, unadulterated by phantasia. The moral of [B] is clear: insofar as there is any such pure perception, perception per se is something distinct from phantasia; perception can and should be independently defined with no reference to phantasia.

However, Aristotle later attenuates the premise that some sentient animals have no phantasia, which undermines [B]. Although De Anima 2.3 balks at the inference from perception to phantasia (414b15-16; 415a10-11),18 this hesitation contradicts a positive testimony in 2.2.19 Later in

17  Aristotle states that by virtue of phantasia ‘a certain phantasma’ which is said not metaphorically occurs in us (428a1-2). Commentators usually take the ‘metaphorical’ use of phantasia/phantasma to refer to an extended use that cov- ers the sense of perception (and even other rational activities involving images), making imaginative experiences the non-metaphorical and non-extended use of phantasia/phantasma (Hamlyn 1968, 131, Hicks 1907, 460, Shields 2016, 281). This statement is thus brought to support a deflationary reading of phantasia which restricts its sense to imagination or imaginative experience. But this can- not be true given Aristotle’s consideration of illusory experiences; worse yet, the fact that Aristotle draws attention to illusory experiences and drops the sense of “visualization” in the rest of the chapter suggests the opposite. That is, the meta- phorical use dismissed here refers back to the imaginative capacity, and Aristotle recommends adopting a broader sense, which is the non-metaphorical use of phantasia he will turn to. See also Frede 2020, 58.

18  “To those living things which have touch, desire belongs as well. But re- garding phantasia things are not clear. One must inquire into that later” (414b15- 16) … “Rather, phantasia does not belong to some, while others live by this alone.” (415a10-11)

19  “For each of the parts [of an insect] has perception and motion with respect De Anima 3.10-11, the premise is rejected by the confession that all animals possess phantasia even though imperfect animals have only indeterminate phantasia.20 The argument runs as follows: animal actions are initiated by desire, and desire is triggered by phantasiai, because animals detect objects of desire by sensorily or rationally representing them (phantasiai) (see also De Motu Animalium 702a15-19). However, what about stationary animals that possess only touch and are called ‘imperfect animals’ (433b31-434a1)?21 Do they also require desire and phantasia, if they do not move at all? Aristotle affirms that they should have them, insofar as they are witnessed to have pleasure and pain as indicators of desire, though the phantasia based only on touch would be non-visual and vague [27]22. Consequently, since all animals have phantasia together with perception, [B] is deficient in establishing the distinction at issue.

[C] Further, perceptions are always true, whereas phantasiai are for the most part false. (428a11-12) Taken literally, the claim that perceptions are always true while phantasiai are mostly false clashes directly with the forthcoming De Anima 3.3, 428b18-30: among three kinds of perception, only proper sensations (sensations proper to each sense organ) can be ‘almost always true’. Among three corresponding kinds of phantasia, despite the other two kinds being fallible, phantasiai from proper sensations are still true whenever their agents, proper sensibles, are present. For example, the sentence ‘it appears (phainetai) green’ is as true as ‘it looks green’ insofar as the green thing is present. “Phantasiai” in [C] might be taken to be imaginative experiences that are always ‘false’, namely, referring to no actual things, but this reading makes [C] a superfluous to place, and if perception, then also phantasia and desire; for wherever there is perception, there is also both pain and pleasure; and wherever these are, of neces- sity there is appetite as well.” (413b21-24)

20  “Insofar as an animal is capable of desire, it is, in virtue of this, capable of moving itself; but it is not capable of desire without phantasia. And all phantasia is either rational or sensory. And in this latter, then, the other animals have a share as well.” (433b27-30) “It is also necessary to consider what initiates motion in imperfectly developed animals, those whose perception is limited to touch; whether or not it is possible for them to have phantasia and appetite. For they appear to have pleasure and pain in them; but if they have these, then it is necessary that they have appetite as well. But how could they have phantasia in them? Or rather, just as they are moved indeterminately, these things are present in them, but present indetermi- nately. Hence, as we said, while sensory phantasia belongs to the other animals [i.e. including imperfect animals], deliberative phantasia belongs to rational ani- mals.” (433b31-434a7)

21  Mobile and sessile animals differ in whether having touch only or having also indirect senses, see 434b22-27.

22  See Papachristou 2013 on the indeterminate phantasia of imperfect ani- mals.

repetition since [A1] and [A2] have already argued the same point. Thus, if we are going to make sense of [C], it should at least spell out what will be confirmed at 428b18-30, that is, the comparison that proper sensations are always true while phantasiai of complex things are mostly fallible (prone to error).23 But this reading is no doubt forced. What is uttered is either an implausible point that contradicts other texts or a repetition of an inadequate point [A1-A2]. [C] cannot justify the distinction at issue either. [D] Again, we do not say that this appears to us to be a man whenever we are in a state of actuality accurately in relation to the sensible object, but rather whenever we do not perceive clearly whether it is true or false. (428a12-15) [26, 28, 29]24 Agreeing with some commentators, I contend that [D] voices the real difference between phantasia and perception [8, 9, 11, 12, 13]25 Unlike [A] and [C], now phantasia is implicated in the explanation of perceptual error that takes place when we sense-perceive instead of imagine, and phantasia’s falsehood is not just “the absence of objects”. [D] claims that whenever we see a man accurately and confidently, we do not say “it appears to be a man” but “we sense-perceive a man”, from which it concludes that perceptions insofar as they are veridical are distinct from fallible experiences (“whether it is true or false”) which are phantasiai. Perceptual error happens, then, whenever we are taken in by fallible perceptions and judge wrongly, confusing “sense-perceiving” with “appearing” and judging mistakenly.

23  Proper sensations are always true because, by definition (De Anima 2.5 and 2.12), they have no way to go astray in accepting the imprint of proper objects (e.g., colors and sounds). As regards proper sensations, we cannot meaningfully speak of the correspondence between contents (aisthemata) and objects (aist- heta) because mismatch is impossible. On the contrary, representation of com- plex things – ordinary things and events – we grasp by predicative perception (aisthesis kata sumbebekos), is prone to error (428b20-21).

24  ἔπειτ᾽ οὐδὲ λέγομεν, ὅταν ἐνεργῶμεν ἀκριβῶς περὶ τὸ αἰσθητόν, ὅτι φαίνεται τοῦτο ἡμῖν ἄνθρωπος, ἀλλὰ μᾶλλον ὅταν μὴ ἐναργῶς αἰσθανώμεθα πότερον ἀληθὴς ἢ ψευδής. Cf. Ross 1956, Hicks 1907, and Jannone 1966 ad loc.: whether it is true or false (ποτερόν ἀληθὴς ἢ ψευδής) (Ross); then it is either true or false (τότε ἢ ἀληθὴς ἢ ψευδής) (Hicks); then both the true and false perception apply (τότε καὶ ἡ ἀληθὴς καὶ ἡ ψευδής) (Jannone). According to Hicks’ and Jan- none’s editions, people are tempted to believe that there is true or false only when phantasia occurs, so perception is always true. Perhaps they intend to make [D] dovetail with [C] which claims “perception is always true”. But taking this pain is gratuitous because Aristotle explicitly says perception is fallible.

25  Modrak 1968, 51-52, 67, Schofield 1992, 260, Turnbull 1994, 320. Cf. Corcilius 2014, 74-75 who, in his rehearsal of the aisthesis-phantasia distinction, unfairly leaves out [D], in order to underline the imaginative aspect of phanta- sia; also Rapp 2001, 93 who, taking the basic sense of phantasia as imaginative ability, blames [D]’s confusion and hesitates to admit the role of phantasia in perceptual error.

“Veridical perception”, as [D] suggests, refers to sensory experiences simultaneously with and corresponding to external sensible objects, especially when it comes to ordinary things.26 “Veridical” implies (1) the accuracy of perceptual contents and the correspondence between contents and objects, as well as (2) the subjective feeling of distinctness and confidence indicative of our access to sensible reality. For instance, I veridically see Hippias if and only if I see it as Hippias, I do not doubt it, and the external thing is Hippias. However, when I am unsure if that is Hippias (“it seems to be him”), my perceptual experience is fallible and a phantasia or phantasma. As confidence is involved in the accuracy of perceptual experiences, veridical perceptions already possess the element of assent (pistis) characteristic of belief (doxa), in terms of which belief is distinguished from non-committal phantasia (428a18-24).27 In contrast, the feeling of uncertainty is a consequence of fallibility. To put it in a nutshell, perception is strictly distinct from phantasia because there are cases in which veridical perceptions, being (1) accurate and (2) assent-bearing, are distinct from fallible and non-committal phantasiai.

Inaccurate perceptions (“it seems to be”) are fallible and deceptive (tempting us to err), but they are not simply false perceptions. As “can be false” does not mean “is false”, inaccurate experiences are fallible instead of false. False perceptions, like veridical perceptions, already contain assent: it is only when I believe my perceptual experiences to be the case that they can be true or false according to the case. By contrast, a fallible phantasia needs not be a false view (doxa), and fallible experiences are not necessarily false. When I glimpse, say, Hippias and say “it seems to be Lysis”, and Hippias is actually over there, I am not thereby asserting a false belief that “it is Lysis”. Phantasia accounts for perceptual error by opening up the possibility of error: our perceptual apparatus may provide us with misleading information in a “fantastic” state of mind, without which there will be no perceptual error, but it is belief, not phantasia, that makes the mistake of accepting this information.28 Hence, it is not false but fallible experiences that are phantasiai and distinguished from veridical perceptions.

Although inaccurate perceptions figure prominently in the account of perceptual error, phantasia in the sense of

26  Although proper sensations are a fortiori veridical by these standards, this notion of veridical perception fits perception of ordinary things better. Whereas proper sensations are in a way, “incapable” of being false, perceptions of or- dinary things imply predicative perception (aisthesis kata sumbebekos) which predicates one item of another and admits of error (428b18-22).

27  That said, perception is arguably not identical with belief, since belief has wider scope and extends to imperceptible things.

28  De Insomniis 460b16-27, 461b4-8.

fallible experience extends beyond inaccurate perceptions. Let me explain. Veridical perceptions by Aristotle’s standards, I argue, exclude imaginative experiences (such as memories, imaginations, dreams, and hallucinations) on the one hand and inaccurate perceptions on the other, so both classes belong to phantasiai due to their fallibility. When I am veridically looking at something, I can neither hesitate over nor imagine this thing at the same time. Obviously, it is impossible that I know I am looking at something and still feel uncertain about whether it exists. But what if I imagine something is the case (e.g., “I think a tree is now in your courtyard”) and guess right? Will it not be veridical? Still, even if I guess right, it is testified afterward. I risk being wrong and cannot be confident when I am imagining. Imaginative experiences cannot be veridical in this strict sense given their fallibility and non-committal attitude, as they can operate independently of external objects ([A1-2] and [C]) and project mental images at will (427b17-24). Veridical perceptions, thus, are neither imaginative experiences nor inaccurate perceptions, both of which are phantasiai in terms of their fallibility.

To sum up, Aristotle introduces a rigorous notion of veridical perception by which phantasia, in the sense of fallible experience, is demarcated. That is, an experience is marked as a fallible phantasia and a period of phantasma through lack of the intrinsic features of a veridical perception; such experiences turn out to be both imaginative experiences and inaccurate perceptions. While the notion of veridical perception safeguards the epistemic reliability of perception, the notion of phantasia makes room for the possibility of perceptual error and thus explains the consistency between perceptual reliability and perceptual error. Besides, this account offers a unified sense of phantasia to reconcile the incompatibility between phantasia as visualization and as sensory appearance, which, as I mentioned at the beginning, is one of the two difficulties in De Anima 3.3. Moreover, this account proves to have abided by MIOP, as the distinction between veridical and fallible experiences clarifies the aisthema-phantasma distinction and the preliminary definition of phantasia (428a1-2) whereby phantasia is defined [9]29.

With this account in mind, we can now better understand Aristotle’s closing remarks on the affinity between perception and phantasia (428b10-429a2). Here, since perceptions of three kinds of objects are fallible to various degrees (428b18-25), from three kinds of perceptual fallibility derive three kinds of fallible experience (428b25-30). In a word, we

29  This account is in principle similar to Modrak 1968, 49: “the awareness of a sensory content under conditions that are not conducive to veridical percep- tion”, but I try to further it by digging into the perceptual-realist framework in the next section.

draw fallible experiences (phantasiai) from sense-perceiving whenever it is unreliable and prone to error (428b14-18). To illustrate, when I am looking at a dim contour of Hippias and hold that “it seems to be Hippias”, I get this ambiguous appearance (fallible phantasia) from my viewing activity in unfavorable viewing conditions (fallible perception) [14]30.

No doubt, the aisthesis-phantasia distinction is in tune with phantasia’s association with perception, and the same concern lies behind the two arguments (428a5-16, 428b10-429a2), which is the harmony between perceptual error and perceptual reliability. The paradox between the aisthesis-phantasia distinction and their association is thus resolved. Perception may fail to access sensible reality (in different ways), and the occurrence of perceptual error rules out such experiences from the class of “perceptions”, i.e. veridical perceptions. On this account, Aristotle is driven to admit a fallible and non-committal kind of experience generated by the sense-perceiving activity, i.e. phantasia, to accommodate perceptual error and keep the reliability of veridical perception intact. Perceptual reliability, therefore, is guaranteed by the notion of veridical perception which is a sign of human beings’ immanent capacity to access sensible reality.

Perceptual Realism

It is already clear that the relation between veridical and non-veridical experiences is asymmetric. Phantasia is characterized and differentiated not by its own inherent characteristics but by what it is to be a veridical perception. To understand the nature of veridical perception, we need to identify the criteria for what qualifies as a veridical perception. I will show in this section that the nature of veridical perception, thereby the distinction between veridical perception and phantasia, presupposes Aristotle’s account of perceptual realism.

Modrak identifies a threefold set of criteria for what sensory experiences are non-veridical phantasiai: (a) an external object is acting upon the percipient but the external conditions are insufficient, e.g. the excessive distance from the object, for a veridical perception to happen; (b) an external object is acting upon the percipient but the internal conditions are insufficient, e.g. being feverish or overly emotional, for a veridical perception to happen; (c) no external object is acting upon the percipient when she is experiencing [9]31.

30  So, there is no oddity as Frede (2020) believes when Aristotle claims that phantasia is an “after-image” of perception, he also pronounces the “simultane- ity” between perception and phantasia (428b25-30).

31  Modrak 1968, 52.

Accordingly, a veridical perception should happen if an external object is acting upon the percipient and both external and internal conditions are sufficient for her to perceive the object exactly as it is. However, what constitutes external and internal sufficiency is not fixed a priori but contextually dependent. In one sense, these conditions could be infinite; in another, being part of these conditions depends on what the object in each case is and what facilitates its cognition. This means that we neither need nor can enumerate these conditions before discerning what it is to have a veridical perception. Instead, identifying veridical perceptions in every case essentially depends on the overarching criterion– what the external object is, and whether it is accurately reflected in perceptual content. At this juncture, we might be tempted by contemporary views, such as phenomenalism and representationalism, to claim that what the object is should instead be determined by both external and internal conditions. Conceivably, we might think that we perceive things only by representing them, and the way we represent things partly determines how they appear to us, whether or not their appearances are themselves per se.32 However, Aristotle is impervious to this sort of challenge if he takes a firm stance on his perceptual realism. Despite recent trends to interpret Aristotle’s perception as ‘representationalism’ [30]33, ‘moderate Protagoreanism’ [31]34, or ‘subtle realism’35, I ally myself with the majority who defend Aristotle’s perceptual realism in this debate [22, 32, 33, 34, 35]36.

In general, perceptual realism claims that what we perceive by senses (perceptual contents) reveals what perceptible things really are in themselves (perceptible objects), independent of our perception. In particular, Aristotle’s perceptual realism asserts that sensible objects exist independently of perception, causally initiate perceptual activity, and partly determine perceptual contents by directly transmitting proper objects (Categories 7, 7b15-8a12; De Anima 2.5 and 12). Indeed, we may think a red thing stops being red if nobody is looking at it. The redness’ power to be perceived as red cannot actualize itself without actually being seen, so perceiving and being perceived share the same actuality, i.e., the same act (De Anima 3.2, 426a2-11). Since perceptible objects derive their essences from their

32  For phenomenalism, what things are in themselves are their appearances; for representationalism, they are still different. But they all deny that reality has an independent existence.

33  Ganson 2020.

34  Gottlieb 1993.

35  Mamodoro 2014.

36  Broadie 1993, Esfeld 2000, Caston 2018, and Gregoric 2019. Caston 2018 provides an overview of this debate.

correlation to perception (Categories 7, 6b35), and actual being-perceived cannot exist without actual perceiving, one might suspect the objectivity of perceptible objects. However, Aristotle claims that the objectivity of perceptible objects is underwritten by the underlying things to which these sensible features adhere: And in general, if in fact only the perceptible object (aistheton) exists, nothing would exist unless living things existed; for there would be no perception. Now, it is doubtless true that [if so] neither perceptible objects nor perceptual contents (aisthemata) (which are an affection of a perceiver) would exist; but that the underlying things (hupokeimena) which produce perception would not exist, even in the absence of perception, is impossible. For perception is not of itself, but there is some other thing too apart from perception, which necessarily prior to perception; for what changes something is prior in nature to the thing changed, and this is no less even if they are called these things correlative to one another. (Metaphysics 4.5, 1010b30-1011a2) Had there never been perceivers and perceptual activities, no perceptual contents of course would have existed; perceptible objects, insofar as they are actually being perceived, e.g. redness actually seen, would have never existed either. However, the red thing would not thereby cease to be red. Its visibility would be unaffected, simply because it, as the underlying material thing, would not thereby disappear when perception disappeared. In other words, although the actualities of perceptible objects and perception are ontologically simultaneous, their potentialities are not (De Anima 3.2, 426a20-25); the perceptibility of perceptible things is ontologically independent owing to their independent material substrata.

Since perceptible objects are real, we have veridical perceptions when we accurately describe these objects, so my perception of something as, say, red is true if and only if this thing is red at the moment. De Anima 2.6 and 3.1 spell out what Aristotle’s sensible objects are. (1) Basic objects are those directly acting upon us (aistheta kath’ hauto), including those proper to each sense organ (ta idia), like colors and sounds, and those common to more than one sense organ (ta koina), like shape and magnitude. (2) Complex objects are constructed from fundamental ones by predication (aistheta kata sumbebekos). For instance, this yellow thing is bitter, and this white thing is Diares. By predication (kata sumbebekos), we perceive sensible compounds of proper and common objects, which reflect ordinary things in the world, such as the yellow, bitter bile, and white Diares.

In contrast to the other two types of sensible objects, proper objects are the most fundamental, for ‘that which psycho-physically moves each sense organ’ forms the strictest sense of ‘sensible object’, from which the other two derive their essences (418a24-25). Common objects are what accompany more than one proper object, and objects by predication (aistheta kata sumbebekos) result from associating proper objects with additional features. Proper sensations are infallible or nearly infallible, whereas the other two kinds are mistakable (428b18-25), which renders proper objects least susceptible to phenomenalist and anti- realist attacks. Aristotle highlights the fundamental status of proper sensations: In the case of our perceptions themselves, the perception of what is alien and proper, or of what is neighboring and what is its own, are not equally authoritative, but in the case of color it is sight, not taste, and in the case of flavor taste, not sight; and each of these never asserts about the same thing in the same time that it is simultaneously so-and-so and not so-and-so (1010b14-19).

For instance, when sight discerns redness, at the very moment it is certain that there is a piece of redness. Since there is no way for this perception to change simultaneously into an awareness of non-redness, it is impossible for the same thing to be red and not red simultaneously, as perceptual anti-realists might assume. At least at the level of proper sensation, perception is a reliable channel for sensible reality, and what it channels, i.e. proper objects, are unquestionably real. Proper sensations, thus, provide clear evidence for the reliability of perception and the reality of sensible objects, even though predicative perception of something belonging to something is still subject to falsity37.

Nonetheless, the hierarchy of three types of sensible objects does not imply a temporal sequence in which our ordinary perception must proceed from the fundamental to the secondary objects. Although Aristotle does not lay down a theory of how we sense-perceive ordinary things like a horse or a landscape, he does underscore the role of common sense (koine dunamis) in sense-perceiving complex objects when answering why one single sense capacity can simultaneously correlate with several proper objects (De Anima 2.1-2, 2.7, 431a17-b1, and De Sensu 7). Presumably, both common and predicative perceptions will be demanded in an act of perceiving, say, the man Cleon together with his action, as they bind different items in different ways [36, 37, 38, 39]38. Our ordinary perceptions, therefore, do not start with

37  “Nor, even in another time, was there dispute about the affection [proper sensations], but only about that in which the affection coincides; I mean for in- stance that the same wine might seem sweet at one time and not sweet at another, if there is a change either in it or in the [human] body; but the sweet such as it is, when it is, has never yet changed, and one always has the truth about it, and anything that is going to be sweet is such of necessity.” (1010b19-30)

38  Scholars stress the role of both common sense (Kahn 1966, Modrak 1981, and Marmodoro 2014) and predicative perception (Cashdollar 1973) in recon- discrete color sensations and build them up temporally, but grasp the entire images of ordinary things at once.

More importantly, this hierarchy should not be misconstrued as implying a sense-datum or idealist theory. It does not suggest that proper objects are objective and the other two kinds are subjective constructions, as if proper sensations alone, being infallible, report reality, whereas the other two kinds merely interpret reality and construct ideas about it. Aristotle’s talk of perceptual truth in De Anima 3.3, 428b18-30 contains no such idea. Instead, from three types of perceptions being differently fallible truth-bearers, we can only infer that there are three kinds of sensible objects acting as truth-makers and yielding perceptions with varying degrees of reliability.39 Hence, external sensible objects include proper objects, common objects, and objects by predication, and all serve as criteria for veridical perceptions.

Aristotle’s own examples of perceptual illusion help illuminate his perceptual realism and the notion of veridical perception. First, a figure may appear to be, say, Hippias from a distance but turn out to be Lysis upon closer inspection (428a14). What individualizes Hippias at the perceptual level is a set of comparatively fixed sensible features: his stature, shape, color, clothing, facial features, sound, etc. These features serve to verify a perception of him. Thus, I veridically perceive Hippias if and only if I experience ‘Hippias’ in Hippias’ physical presence, that is, if and only if some or all of his sensible features are present in my perceptual contents while he is present. In reverse, if I somehow experience Lysis’ features in Hippias’ presence, I have a deceptive phantasia. Second, the sun appears to be one foot in diameter but is actually larger than the Earth (428b2). Due to the foreshortening effects, the perceptible size of a thing may be deceptive. Still, the sun has its real size, even if its accurate perception is practically impossible. As long as my perception of the sun’s size (“one foot”) fails to reflect its actual magnitude, my perception of this “common object” is an illusion. Third, the land may appear to be moving past me as I sail past it (De Insomniis 460b26). The appearance of the passing land is illusory since the land, of course, remains stationary while I am in motion, and the relative movement between the boat and the land distorts my perception. My perception of the “object by predication” (i.e., “the land is moving”), is deceived under these adverse conditions. These examples clarify Aristotle’s principle: a structing Aristotle’s theory of ordinary perception. Note that by ‘Cleon’ (425a25) and ‘Diares’ (418a21) Aristotle does not mean the substantial form of Diares but only the sensible compound or the compound of sensible forms we perceive and designate ‘Diares’.

39  According to Aristotle’s correspondence theory of truth, cognition is true only by corresponding to reality.

perception is veridical and different from a phantasia if it corresponds accurately to perceptible reality, which consists of three kinds of sensible objects. These objects constitute reality and provide the standard of truth in sensory experience. Besides, the case study illustrates Aristotle’s point that perceptions of common objects and objects by predication are significantly more prone to error than those of proper objects, with the former being the most fallible (428b19-25). Our ability to discern sensible compounds and events by predicative perception usually fails because we misperceive the common objects within.40

Conclusion

Traditional interpretations of Aristotle’s concept of phantasia and De Anima 3.3 are ineffective in addressing two difficulties: (1) the paradoxical relation between perception and phantasia; (2) the incompatibility between phantasia in the sense of imaginative experience and appearance. The interpretation I offer seeks to overcome these two difficulties by focusing on the aisthesis-phantasia distinction and looking at this distinction as a response to the challenge of perceptual unreliability. I claim that this distinction arises from Aristotle’s need to explain the possibility of perceptual error while upholding the reliability of perception. It lies between veridical perception and non-veridical phantasia and rooted in Aristotle’s perceptual realism. Aristotle does not justify perceptual realism through this distinction but rather builds this distinction on the foundation of perceptual realism. This account of phantasia provides a possible way to resolve the difficulties at issue: (1) phantasia in the sense of fallible experience unifies both imaginative experience and inaccurate perceptions; (2) phantasia’s distinction from and association with perception are both drawn in answer to the challenge of perceptual unreliability.

As fallible phantasia safeguards the epistemic reliability and the cognitive value of veridical perception, this phantasia is itself non-cognitive: we are not expected to discern reality by this unreliable means. However, this notion still faces the challenge of having to explain the positive role of phantasia in Aristotle’s doctrine of concept formation (Metaphysics 1.1, Posteriors Analytics 2.19, De Anima 3.8), which demands a notion of cognitive phantasia. I suggest it is rather the task of

40  However, there is a controvertible case. Two crossed fingers appear to be two to touch but one to sight (Metaphysics 1011a34; De Insomniis 460b20). Among proper sensations, sight is more authoritative than touch so visual grasps of common objects are more privileged than tactile grasps to claim truth, to the effect that the visual pronunciation that ‘it is one bundle’ demotes the tactile pronunciation that ‘there are two things’ to a mere illusion. But this visual supe- riority can be accepted only if we prefer one description to another, forcing the fact to be ‘it is one bundle’.

rescuing perceptual reliability from the phenomenalist and skepticist accusations that cries for a notion of non-veridical phantasia in De Anima 3.3, for which reason a notion of cognitive phantasia has to be left undeveloped there. But this task cannot be the only one for Aristotle’s phantasia to fulfil.