1 Introduction

In De divinatione per somnum, Aristotle explicitly rejects the idea that a deity might send us divinatory dreams¹ in order to inform us of things that will happen to us in the future.² In the Arabic tradition, however, Aristotle’s name is linked to a theory that strives to explain veridical dreams on exactly such a basis – as has been pointed out already in the previous chapter.³ While the firm place assigned to prophetic dreams within the religion of Islam (as well as Judaism and Christianity) and their general acceptance as a regular fact of life within medieval Muslim society⁴ will have contributed to the positive reception and further development of this theory among philosophers such as al-Fārābī, Avicenna,⁵ and Ibn Bājja, its origins – and those of its attribution to Aristotle – lie in the Arabic adaptation of Aristotle’s Parva naturalia (Kitāb al-Ḥiss wa-l-maḥsūs).⁶ In this adaptation, the original Aristotelian text, present only in often-distorted fragments, fades into the background in favour of more Neoplatonic and Galenic ideas, especially where the topics of memory and dreaming are concerned. It is therefore not really surprising that we find such a theory of divinatory dreams even in the work of a staunch Aristotelian like Averroes (Ibn Rushd, 1126–98). His thoughts on divinatory dreams are expressed in his Explanatory Paraphrase of Aristotle’s Parva naturalia (Talkhīṣ Kitāb al-Ḥiss wa-l-maḥsūs, completed in 1170)⁷ – that is, in a text based on the very adaptation in which we find the deviant theory of veridical dreaming presented in Aristotle’s name.

This may be a quite straightforward explanation for Averroes’ uncharacteristic divergence from Aristotelian doctrine. It does not, however, resolve the puzzle entirely: it still seems to call for an explanation that Averroes, of all people, should not have blinked at the ideas transmitted in the Parva naturalia adaptation, which, after all, are not exactly similar to those he would have encountered in other Aristotelian works. In this contribution, I want to take a closer look at some aspects of Averroes’ thoughts on divinatory dreaming in comparison with Kitāb al-Ḥiss and to investigate exactly how he responds to his source text. As I hope to show, in the manner in which he interprets and reshapes the doctrine of divinatory dreaming that he finds in the pseudo-Aristotelian text he demonstrates himself once more to be a committed Aristotelian.

2 The Source: Kitāb al-Ḥiss wa-l-maḥsūs and Its Theory of Divinatory Dreaming

We do not know the names of the people involved in translating and adapting the Parva naturalia; in all likelihood, though, the adaptation was produced in the ninth century in the circle of al-Kindī, the group of translators responsible for the Arabic Plotinus and Proclus.⁸ The most striking characteristic of this adaptation, and the most important thing to note in our context, is that only a relatively small proportion of the text actually goes back to the Parva naturalia at all; and even in passages which do reflect the Aristotelian text, preserving more or less coherent pieces of translation, this does not mean that its philosophical content is in any way accurately represented. What the adaptation conveys in terms of doctrine therefore is, on the whole, rather different from what we find in the original Parva naturalia.⁹

Thus the most significant feature of the adaptation’s account of divinatory dreaming (as well as of memory) is a theory of three “spiritual faculties.”¹⁰ These are three post-sensatory faculties located in the three ventricles of the brain, the imaginative or formative faculty (muṣawwir), the faculty of thought (fikr), and the faculty of memory (dhikr). These faculties process the sense perceptions that have been perceived by the five external senses and collected by the common sense. What is particularly remarkable about the way the adaptation describes their respective functions is that it conceives of them as consecutive stages of a process of purification that, step by step, removes all corporeal aspects of a sense perception. The first, corporeal stage is the perceptible object itself; the second stage is achieved with sense perception, which separates the perceptible form from the object; this form is then passed on to the formative faculty, which retains it in the absence of the sense object and passes it on further to the faculty of thought. Stripping away the remaining corporeal aspects (at times described as “shells” or “crusts”), this faculty then distinguishes between the form and its maʿnā (the “core”). This term seems to stand for something like the cognitive content of the perception, the thing in so far as it is purely thought (rather than imagined).¹¹ At the last, fifth stage this maʿnā is passed on to the faculty of memory, which stores it. The maʿnā is entirely spiritual, as no corporeal aspect is left at this point in the process. Nevertheless it is not a universal concept; remaining tied to the original perception, it retains its particular character.

When the adaptation calls the three post-sensatory faculties (or their objects) “spiritual,” this goes back, in the first instance, to their roots in medical theory: located in the ventricles of the brain, they run on pneuma or spirit (rūḥ). However, the anonymous adaptor of the Parva naturalia does not stick to this (lastly material) concept of spirit. In a move that we also know from other texts related to the Kindī circle,¹² he presents spirituality as the opposite of corporeality, aiming for a strict dichotomy between things belonging to the spiritual realm and things belonging to the corporeal realm (see pp. 117–19 below). Nevertheless the fact remains that the “spiritual” faculties (which, after all, belong to the animal soul) and their objects are somehow to be distinguished from the intellectual and divine. As a result the concept of spirituality vacillates between something that is situated in between corporeality and incorporeality, and something that is equated with incorporeality. This tension is observable in particular in the context of the discussion of divinatory dreaming, where the text on the one hand strives to emphasise the sublimity of the phenomenon, while on the other maintaining that it is bound to the “spiritual” faculties of the animal soul.

This is because in dreaming it is again the same three faculties that play the leading roles.¹³ During sleep they are not confronted by any fresh external perceptions that would require their attention; this gives them the freedom to occupy themselves with the stored forms and maʿānī resulting from prior perceptions. Veridical dreams that predict future events, however, cannot originate in such earlier perceptions. Instead Kitāb al-Ḥiss stipulates that in their case, the “Universal Intellect” conveys the dream-image, the “spiritual form” constituting the dream, to the faculties of the dreamer. It also conveys the corresponding maʿnā, that is, the knowledge of what the dream signifies, to the dream interpreter. Since these two and the actual corporeal form that will at some later point appear in the world are related to the same “intellectual form” and hence to each other, dream, interpretation and the actual future event will correspond to each other. The ultimate cause behind all this is, however, God:

[T₁] The sound, spiritual dream-vision is the one which occurs from intelligibles of the Universal Intellect, not from intelligibles of the acquired intellect, [i.e., it comes from intelligibles] which are unknown to the common sense and have not been imagined by the formative [faculty]; the ma‘nā of which [the faculty of] thought does not know, and which are not deposited in [the faculty of] memory. […]

This [kind of] true dream-vision, the cause and reason of which is the true Deity, great be His praise, occurs through the mediation of the Intellect. For whatever the Deity, great be His praise, wanted to become manifest in this world He gave form to in the Intellect at one stroke, and gave form to its forms in this world at one stroke, together with what they imply rationally. The Intellect then made [them] manifest to the soul and to each one of its faculties, according to the measure in which the soul decided that [each] faculty could receive [them]; with the Supreme Cause, I mean the Deity, great be His praise, having created [them] in this way, when He created the Intellect at that time, in order to make manifest what is within it; because the Deity moved [the intellect] at that time in order to make manifest what is in it.¹⁴

Obviously, this theory has nothing at all to do with what Aristotle says in De divinatione per somnum. In contrast to us, however, Averroes did not have the advantage of being able to read Aristotle’s Greek original. When he composed his paraphrase of Kitāb al-̵Ḥiss wa-l-maḥsūs, he seems to have done so on the assumption that the work was by Aristotle; at least, he never expresses any doubt in this respect. He explicitly refers to Aristotle as the author of the text, not just in his paraphrase of Kitāb al-Ḥiss itself, but also, twenty years later,¹⁵ within his Long Commentary on De anima.¹⁶ Moreover, in The Incoherence of the Incoherence (1180), Averroes uses the text as his reference point for true Aristotelian doctrine when he attacks Avicenna’s system of five internal senses;¹⁷ and he generally models his own epistemological psychology more closely than the latter on the theory of the “spiritual faculties” found in Kitāb al-Ḥiss.¹⁸

This does not mean that Averroes was unaware of the vagaries of translation and textual transmission. For instance, he notes that Kitāb al-Ḥiss does not transmit all treatises of the Parva naturalia;¹⁹ and in his Long Commentary on Aristotle’s De anima, he occasionally makes use of a second Arabic translation of the source text wherever he thinks that its readings can help to clarify matters.²⁰ Given his awareness of problems associated with translation and transmission, he may well have entertained doubts as to certain details of the theory propagated by Kitāb al-Ḥiss, even if, as we may have to assume, he accepted its general tenets as part of Aristotelian doctrine.²¹ In the remaining sections of this chapter, we will encounter several instances where he did not in fact follow his source text on every point.

3 Veridical Dreaming as “Potential Sense Perception”

The Arabic adaptation of the Parva naturalia is structured into three treatises or maqālāt: while the first and the last contain the equivalents of De sensu and De longitudine et brevitate vitae respectively, the second maqāla comprises the equivalent of De memoria on the one hand, and a “Chapter on Sleep and Waking” on the other, in which the topics of the three Aristotelian treatises De somno et vigilia, De insomniis, and De divinatione per somnum are dealt with together.²² The arrangement is the same in Averroes’ Explanatory Paraphrase.²³ Furthermore, in both texts the topic of divinatory dreaming is not confined to the last part of the treatise, but is present from the start. Thus, our first pair of textual examples is taken from the beginning of the section on sleep. The passage from Kitāb al-Ḥiss, which will be quoted first, reflects a few lines from the first chapter of De somno et vigilia.²⁴ Italics indicate a relation, however tenuous, to the Greek text:

[T_2a] (1) […] Therefore, the privation [or: absence] (ʿadam) of sleep is waking. This can be verified and recognised when [we consider] the waking and the sleeping person. For a sleeping person will perceive many things while having no doubt that those things he is perceiving in his sleep are there [perceived by him in reality] in his waking state. (2) The difference between the perception of the waking and that of the sleeping person is that the sleeper perceives internally only – and that [kind of] perception of his [takes place] without any movement on his part – whereas the waking person perceives externally, that [kind of] perception [taking place] through movement. (3) Let us therefore say: the difference between the sense perception of the sleeping and that of the waking person has become clear, [i.e.] which one of the two [types of] sense perception is more apt and more correct; this ought to be recognised and known. (4) Let us therefore say: the sense perception of the sleeper is sense perception potentially, whereas the sense perception of the waking person is sense perception actually. (5) Whatever is potential is hard to perceive, whereas what is actual is perceptible and can be known. (6) However, the sleeper’s sense perception, even though potential, may well emerge into actuality; although some of it will emerge in a clear and plain manner, while some of it will be difficult and unclear. (7) As for [the question of which one is] the most perfect and the noblest of the two: the spiritual is nobler than the corporeal. However, the spiritual is not considered nobler than the corporeal by the corporeal, nor is the corporeal considered nobler than the spiritual by the spiritual; rather, the spiritual is considered nobler than the corporeal by the spiritual, whereas the corporeal is considered nobler than the spiritual by the corporeal; yet it is not at all possible that the spiritual should be considered nobler by the corporeal, whereas it may indeed be possible that the spiritual, which we have said to be potential, is considered nobler by man than the corporeal, which we have said to be actual. (8) Evidence for the spiritual being nobler than the corporeal is that the spiritual indicates what will come to be in the future, whereas the corporeal only indicates what has come to exist at the present time. (9) When a person unites his faculties through the subtlest of things and makes them one, he will see the very thing he is seeing potentially in the same way as someone would see it actually. It is just because his faculties are separated that a man is prevented from seeing things in potentiality in the same way as the things he sees in actuality […].²⁵

Elsewhere,²⁶ I have demonstrated in detail how in this passage the adaptor uses a string of keywords taken from the Greek text (note the words in italics) in order to create a strict dichotomy according to which sleep is associated, among other things, with the “internal” perception of the future (i.e., veridical dreams) and with spirituality, whereas the waking state is associated with “external” perception of the present and with corporeality – which renders it less “noble” than the state of sleep. Here, I want to focus on one particular aspect of this dichotomy: the association of the sleeping state (and of veridical dreaming) with “potential sense perception” (4, 6).

In the corresponding passage of the original Parva naturalia, Aristotle explains sleep with reference to waking – it is, as it were, the absence or privation²⁷ of waking – and to sense perception: waking is characterised by sense perception, sleep by the absence of it.²⁸ Hence, sleep is an affection of the perceptive part and applies to all beings with sense perception, i.e., to all animals.²⁹ While the Arabic adaptation preserves aspects of this idea, there are stark and crucial differences. In our passage it is not the case (as in Aristotle) that during waking, both internal and external perceptions are taking place, whereas they are absent during sleep; instead the adaptor distinguishes between two different kinds of perception, “external” versus “internal” perception (2). Internal perception during sleep constitutes potential perception, external perception during waking actual perception (4). The second half of this latter claim seems quite innocuous, but what does the adaptor mean by talking of a sleeper’s sense perception as “potential”? In the parallel passage in De somno et vigilia, potentiality (or capacity) and actuality (dýnamis/enérgeia) are employed within a context where Aristotle argues that sleep and waking are common to body and soul: if waking and sleep are characterised in terms of the presence or absence of perception, then, given that perception affects both body and soul, both these states must belong to body and soul, too: waking as the state in which the perceptive capacity is actualised, and sleep as the state in which this capacity of the soul is present but not actualised (and, in fact, temporarily inhibited).³⁰ This argument is not taken up in the passage from Kitāb al-Ḥiss,³¹ which here merely picks out the keywords dýnamis/enérgeia. It furthermore seems clear that Kitāb al-Ḥiss does not refer to sleep, or to “potential perception,” in the sense of a merely unactualised or temporarily suspended capacity of sense perception. “Potential perception” is rather set up positively as a different, alternative type of perception. It takes place during sleep and without movement, is “internal,” more difficult, and has different objects from regular perception taking place during waking, which themselves are described as “potential” in contrast to the “actual” objects of perception (5). It is linked to “spirituality” (7), and as the adaptation explains further on, its objects are future things or events, which have not yet materialised in the corporeal world (8). This is a clear reference to veridical dreams,³² which the adaptor thus prominently introduces into this passage at the beginning of the chapter.

In a broad sense this idea might yet be seen as derived from, or at least not entirely alien to, Aristotelian notions of sleep and dream. After all, in De anima sleep is linked to the potentiality of perception in so far as Aristotle uses sleep in order to illustrate that perception is spoken of in actuality as well as in potentiality.³³ On the other hand, Aristotle also speaks of dreaming as some kind of perception:³⁴ dreams, though not constituting perception in an unqualified way,³⁵ belong to the perceptual part (in its imaginative function);³⁶ they are appearances (phantásmata) based on affections previously produced by sense objects, which linger on in the sense organs when the perception is no longer actualised.³⁷ However, explaining the notion of “potential perception” that we find in Kitāb al-Ḥiss as a somewhat rough-and-ready combination of these two aspects of Aristotelian doctrine fails to capture some of the more significant points of the idea developed by the adaptor in our passage, in which he clearly deviates from Aristotle. To start with, the text here refers to predictive dreams only, i.e., to dreams that in fact do not go back to previous perceptions³⁸ as described in De insomniis.³⁹ This raises the question of whether the adaptor regards the term “perception” as implying that something constituting a perceptible object is in fact being perceived, which would not be the case in an ordinary dream. Ordinary dreams would then be missing from the account simply because they do not belong to the class of things the text is discussing in the passage. However, the description in T_2a (1), which is supposed to justify labelling dreams as “perceptions,” fits both ordinary and veridical dreams. Furthermore, where the existence of non-veridical (“vain”) dreams is acknowledged in other parts of Kitāb al-Ḥiss, they are explained in an exactly analogous way to veridical dreams, the only difference being their source (i.e., earlier perceptions as opposed to forms conveyed by the Universal Intellect).⁴⁰ Nowhere does Kitāb al-Ḥiss claim that veridical dreams are perceptions while other dreams are not; and ordinary dreams are described in the language of perception (especially that of seeing), just as veridical dreams.⁴¹ It hence would perhaps go too far to assume that the adaptor holds a positive, explicit theory about ordinary dreams not counting as perceptions. It seems more likely that he simply ignores non-veridical dreams in this passage because they do not fit the strict dichotomous approach he is following here, contrasting perception of the present during waking with perception of the future occurring during sleep. In this system there is no room for ordinary dreams, even if they were excluded from the label “perception”: it would still upset the adaptor’s neat dichotomy to have to explain another perception-like phenomenon that also happens during sleep. However, this dichotomy not only means that perception during sleep is identified with veridical dreaming; the adaptor also spends much ink on driving home the point that this “potential” perception during sleep is “spiritual” and hence “nobler” than actual, “corporeal” perception (7). This fits in with another important aspect of the passage: the thought that sleep, by virtue of the absence of perception, constitutes a privation of the waking state – central to the point Aristotle is making in the corresponding passage – is being lost. Not only does potential perception seem to denote an actually occurring activity (i.e., veridical dreaming) rather than the mere potentiality or capacity for such an activity; in (1), waking is furthermore explicitly described as the absence of sleep rather than the other way round. This is not a slip of the pen: the potential is indeed considered primary (“nobler”) in comparison with the actual (7), again a clear break with Aristotelian doctrine.⁴² Here the text is reminiscent of the idea of a “potency higher than act” which the Arabic Plotinus claims for the intellectual realm:⁴³ in order to perform their acts of intellection, intellectual substances do not need to transfer from potentiality to actuality. That is only necessary for perception in the corporeal world – where things are wrapped in “shells” that first have to be penetrated.⁴⁴ There are enough points of contact with Kitāb al-Ḥiss – the involvement of the “Universal Intellect” in veridical dreaming, the adaptor’s emphasis on its sublimity and “spirituality” in comparison with “corporeal” perception, and, of course, the notion of the “shells” that have to be stripped away – for us to assume that this concept of a “potency higher than act” may very well have been looming in the back of the adaptor’s mind when working on the passage quoted above. This is all the more plausible given that, as mentioned above, Kitāb al-Ḥiss generally shares some characteristic traits with other texts produced in the circle of al-Kindī, and in particular with the Arabic Plotinus.⁴⁵ Nevertheless, it is not a complete fit: after all, the “potential perception” at issue here, that is, veridical dreaming, is still an activity carried out by the “spiritual faculties” rather than the (Universal) Intellect, and certainly is not something that would be going on incessantly. Nor is it completely detached from the Aristotelian concept of potentiality. Thus we read in (6) that “the sense perception of the sleeper, even though being potential, may emerge into actuality; although some of it will emerge in a clear and plain manner, while some of it will be difficult and unclear.”

How are we to understand this sentence? What exactly is supposed to be actualised here? According to one possible option the remark would refer to the fact that in veridical dreaming the dreamer perceives something that is still potential but is guaranteed to become actual in the future, when it will also be actually perceived by the person who is now dreaming about it. It would hence focus on the potentiality/actuality of the perceived object; this would go well with clause (5), which in fact addresses the potentiality and actuality of objects of perception. However, on this reading it is problematic that in (6) the adaptor speaks of the activity of sense perception (ḥiss) rather than of the perceived and its emergence from potentiality into actuality. Hence one may ponder another interpretative option: that the text is referring here to the activity of veridical dreaming (i.e., of perceiving the form of a future perceptible), which may or may not become actualised during sleep. In any case, the idea that there may be some difficulty or lack of clarity connected with the actualisation (as suggested in (6)) could more easily be explained as signifying that a dream may be more or less clear (i.e., as a symbolic representation of the future event), than having it refer, as in the first interpretation, to the possibility that there might be some obscurity connected with the actual occurrence of the event in question or that it might be difficult to relate it back to the corresponding dream.

On this interpretation, (6) would therefore refer to potential potential sense perception as turning into actual potential sense perception – suggesting that the adaptor perhaps got somewhat entangled in his own notion of potential sense perception, a notion that vacillates between an unactualised capacity of sense perception and a higher form of perception in its own right. This does not, incidentally, need to invalidate or exclude the first interpretative option. In fact, the status of the object as still “potential” (in terms of its existence in the world) might yet have played a part in the adaptor’s thoughts: if the object responsible for the actualisation of the perception is itself not actual but still in potentiality, how could the resulting activity turn out to be anything but “potential”? Sentence (6) may well be an attempt on the adaptor’s part to reaffirm the thought that in a veridical dream we are indeed having some kind of perception or perceptual awareness; this would also accord with the spirit of (9), where it is emphasised that such perceptions are on a par with and as valid as “actual” perceptions during waking. However, even here the adaptor balks at recognising the dream itself as something actual (in the sense in which dream images, once they are seen, would be actual in an Aristotelian reading): having a veridical dream only means seeing something “in the same way” as one would see it in actuality, i.e., if it were a regular sense perception.⁴⁶ Actuality proper is reserved for the perception of the sensible object that has actual existence in the outside world.

How, then, does Averroes deal with this passage and the tension it creates for the concept of potentiality? In the parallel passage of his Explanatory Paraphrase, correspondences can be clearly identified (and are indicated by the numbers referring to various phrases of T_2a), but there are also significant departures from the source text:

[T_2b] (1, 4a) We say that sleep and waking can be given various descriptive accounts (rusūm). One of them is that sleep is potential sense perception. For it is apparent that the sleeper sees [or: believes] [in his dream] that he is eating and drinking and perceiving with all his five senses. Waking, however, is actual sense perception. From these two descriptive accounts it emerges clearly that sleep is the privation of waking. For what is potential is the privation of what is actual; (6, 8) but the sense perception that is potential during sleep may happen to emerge into actuality; this occurs in the case of true dreams and miraculous warnings. (7a) In these cases potential sense perception is nobler than actual sense perception. False potential sense perception, however, is lowly; the one that is actual is nobler than it. (4b) It seems to be the case, as Aristotle says, that actual perception is corporeal and potential perception spiritual. (7b) The corporeal is nobler for the corporeal perceiver, whereas the spiritual is nobler for those who are spiritual perceivers. The spiritual is not nobler than the corporeal in the eyes of the corporeal, nor is the corporeal nobler in the eyes of the spiritual perceiver. In an absolute sense, the spiritual is nobler than the corporeal. (9) Spiritual perception is not only found during sleep alone, but may also be found during waking when the three faculties are joint and united, as has been said before. (10) From these two descriptive accounts it emerges that these two potencies [i.e., sleep and waking] are one with respect to their subject, but two in essence and definition; their location is the sensually perceiving faculty, and they are common to soul and body.⁴⁷

Rather than passing over the problematic notion of “potential sense perception” (given that it does not play much of a role in the rest of Kitāb al-Ḥiss, this would have been a possible option to consider), Averroes focuses directly on the claim that sleep is potential perception and waking actual perception (1). He neglects the point about internal and external perception made in Kitāb al-Ḥiss (2), and does not follow the adaptation in assuming that perception during sleep is generally “more apt and more correct” (3), nor in emphasising that “whatever is potential is hard to perceive, whereas what is actual is perceptible and can be known” (5). With all these points the adaptor suggests that there are two kinds of perception, one linked to sleep, one to waking, which can be distinguished by their objects (the spiritual – the future; the corporeal – the present), employ different methods (and “organs”: internal vs. external senses), and can be compared with each other in terms of their “correctness.” Averroes, by contrast, emphasises a more Aristotelian understanding of the relation between sleep and waking as well as between potentiality and actuality: “sleep is the privation of waking. For what is potential is the privation of what is actual” (1). It may not be unproblematic to describe the potential as the privation of the actual⁴⁸ (and Averroes has in fact a much more nuanced understanding of potentiality and actuality).⁴⁹ However, faced with the claim of his source text that waking is the privation of sleep (T_2a (1)), Averroes here clearly feels the need to rectify this, not just by returning to the reverse, original Aristotelian position, but also by supporting it argumentatively in pointing out that (if anything) it is potentiality that is the defective state, not actuality. In doing so, he also corrects the claim of Kitāb al-Ḥiss that the potential takes priority over the actual (T_2a (7)).

Nevertheless, Averroes does not quite confine himself to this understanding of potentiality and actuality, according to which potential perception would mean merely the absence of all perception coupled with the possibility that it will occur when the sleeper awakes. Consider the third sentence of our passage. Where we would expect a straightforward explanation of why sleep is the privation of waking, we read instead: “For it is apparent that the sleeper sees [in his dream] that he is eating and drinking and perceiving with all his five senses.” This seems to indicate that there must be more to “potential perception during sleep” than a mere lack of sense perception. Some activity is going on that somehow is a candidate for being described positively as “perception.”

However, the continuation of the sentence (parallel to the difficult clause no. 6 in the adaptation) shows that for Averroes this does not preclude but rather reaffirms the understanding of sleep as a state of unactualised sense perception: in the case of veridical dreams, “the sense perception that is potential during sleep may happen to emerge into actuality.”⁵⁰ While during sleep I see something only as a dream, I will perceive it actually in the waking state. Here the difference between potential and actual appears indeed to attach to the act of sense perception rather than to its object, even if the act’s emerging into actuality also naturally presupposes the presence of the actually existing object.⁵¹

By defining potential sense perception in relation to actual perception, as perceiving something in a dream that one may or may not perceive later on in actual reality, Averroes also leaves room for ordinary dreams in his account: they are the (acts of) perception during sleep that will not emerge into actuality (in the sense of becoming actual in waking life). While this move makes the account more convincing than that of Kitāb al-Ḥiss in this respect, it also means that perception during sleep cannot per se be called “noble” except in the case of veridical dreams (in fact, false, deceptive dreams rank lower than actual sense perception during waking, (7a)). However, this then must also apply to the epithet of spirituality, which is closely linked to that of “nobility.” Thus, potential perception cannot simply be equated with spiritual perception, as the adaptor of Kitāb al-Ḥiss would have it. Perhaps this is one reason why Averroes expresses himself somewhat carefully on this last point: “It seems to be the case, as Aristotle says, that actual perception is corporeal and potential perception spiritual” (4b). For in Averroes’ account, “potential perception” covers true and false dream-visions, whereas the attribute “spiritual” is preserved for true dream-visions only. This can be gathered from the information that “spiritual perception” may also occur during waking (9), a reference to prophetic visions, as can be gleaned from another passage of the text⁵² (where, incidentally, the notion of potentiality does not play any role).

Averroes here risks disagreeing with (ps.-)“Aristotle” for the sake of maintaining the basic Aristotelian concept of potentiality and actuality. This also enables him immediately to get to the point that is in fact at stake in the passage from De somno et vigilia: sleep and waking both affect the faculty of sense perception, and therefore belong both to soul and body.⁵³ Rather than following Kitāb al-Ḥiss in establishing a second, metaphysically loaded, “spiritual” kind of perception, Averroes insists that potential perception during sleep is related to actual perception during waking in the usual way, as it were: in the case of veridical dreaming, potential perception has an equivalent actual perception in the waking state.

What about the mental activity taking place during dreaming, which has been labelled “potential perception” but which is more than just the absence of actual sense perception (1)? Taking the five-stage process “from perception to maʿnā” described in Kitāb al-Ḥiss (see section two above) as his starting point, Averroes describes dreams as reversed perceptions:

[T₃] (1) We shall say that since the sleeper senses as if he were seeing, hearing, smelling, tasting, or touching while there are no external sense objects present, it must follow that in sleep this movement must have its origin where it ends up in waking. (2) And since in waking this movement originates with the external sense objects and finally ends up at the faculty of memory (which is the fifth stage), it follows that its origin must lie with that faculty – (3) except that the only faculty active during sleep is the imaginative one, since the faculties of thought and memory are not active during sleep. For this [imaginative] faculty is in permanent movement and unceasingly active conceptualising, creating similes, and transitioning from one image to the other […] (4) From all this it is clear that, of all faculties of the soul, dream-vision, whether true or false, is primarily related to the imaginative faculty.⁵⁴

During the regular perceptive process, the senses are affected (“moved”) by the perceptible objects; this “movement” is then passed on to the common sense, the imaginative faculty, and so forth (2). In dreaming, however, the process runs in the reverse direction. It is not a strict reversal, though, as Averroes has to admit: rather than the faculty of memory, it is the imaginative faculty that drives the process, as this is the only faculty that remains active during sleep (3).⁵⁵ As Averroes goes on to explain, it takes a maʿnā of some previously perceived thing and represents it as an imaginative form, as the perceived form of a perceptible object. It thus “moves” the common sense faculty, which in turn “moves” the senses, so that the sleeper will get the impression that he is perceiving something with his senses.⁵⁶ This is possible in sleep and also, during waking, in states of fear or illness because the cogitative faculty is not active and has ceased to control the imaginative faculty.⁵⁷ Also, there is no constant stream of new perceptions to be processed by the imagination during sleep.

In this account of dreaming as reverse sense perception Averroes goes beyond his source text, which contains some of its elements – for example, the idea that the post-sensatory faculties are left freely to pursue their own actions during sleep, and that the “formative faculty” presents dream images to the common sense faculty – but does not string them together in the same systematic fashion. However, Averroes may have found inspiration in the works of al-Fārābī and Avicenna, who conceive of the dreaming process in similar ways.⁵⁸ One aspect in which Averroes’ account (as well as those of his predecessors) departs from Kitāb al-Ḥiss is the thought that the imaginative faculty alone stays active during sleep (3); in Kitāb al-Ḥiss the faculties of thought and memory remain active as well as the formative faculty. Curiously, all three thinkers share the adaptation’s assumption that the common sense also continues to be active, in its function as “screen” on which the dream images are played out – a striking development considering that Aristotle argues, in De somno et vigilia, that sleep must affect the primary perceptive part.⁵⁹

For Averroes, the activity of dreaming is therefore an activity of the imaginative faculty: “dream-vision, whether true or false, is primarily related to the imaginative faculty” (4). The question whether it is, as such, actual or potential sense perception therefore seems to be misplaced. How, then, could Averroes justify his going along with his source in applying the label of “potential sense perception” to dreaming? First of all, Averroes may have regarded it as trivially true that every dreamer, qua sleeper, is in a state of potential sense perception. In thinking, beyond that, of the dreaming activity as “potential sense perception,” two points may have made this move at least tolerable for Averroes: even if dreaming is primarily a matter for the imaginative faculty, the common sense faculty is still involved, albeit in an “inverted” fashion. Furthermore, the sense faculty’s state of potentiality is, from this perspective, being considered in relation to specific, particular acts of perception concerned with specific objects. It is easy to see why this is an attractive move in the context of veridical dreaming, since it allows to formulate a relation between dream and future event: what is seen in a dream while it is as yet potential will later be seen in actuality. As indicated by the last phrase, this relation, however, only works via a reference to the object of perception.⁶⁰

Averroes, then, does in the end leave firm Aristotelian ground with his take on dreams as potential sense perception. However, considering the starting position provided by Kitāb al-Ḥiss I think it is fair to say that he does much to “re-Aristotelianise” the concept. It is also worth noting that Averroes’ misgivings about the idea seem to have won the day in the end: in both his Middle and Long Commentary on De anima, written in the early 1180s,⁶¹ he gives up on applying the label “potential perception” to dreaming, explaining instead that it is neither perception in act nor in potency:

[T₄] There is a sort of imagination which is not sensation in act or in potency, namely, the imagination which comes about in sleep. For it is evident that the imagination which is in sleep, insofar as it is in act, is not sensation in potency, and insofar as that act belongs to it without the presence of the sensible things, [imagination] is also not sensation in act.⁶²

4 Veridical Dreaming as Acquisition of Knowledge

Characterising veridical dreaming as a kind of sense perception has a certain immediate plausibility to it, not just because of the perceptive quality that dreams have when we experience them, but also because veridical dreams are supposed to concern particular (future) events,⁶³ that is, things that we would normally access through sense perception. A perhaps more general and usual way to capture what veridical dreams are and do is to say that they provide the dreamer with knowledge about the future. What makes them so special is, of course, that this is a sort of knowledge we are not capable of achieving on our own, exactly because we cannot perceive things that are as yet non-existing. While traditional accounts of veridical dreaming content themselves with gesturing towards a deity or other supernatural figure that simply sends such dreams to selected people, Kitāb al-Ḥiss undertakes to give a more ambitious philosophical explanation of how this would actually work.⁶⁴ Nevertheless it retains a strong sense of the supernatural, divine, and unaccountable character of veridical dreams, emphasising the gap that exists between the “spiritual” knowledge they supply and the ordinary type of information we access through our senses and faculties.

Averroes is apparently not entirely satisfied with the explanation offered by Kitāb al-Ḥiss, especially when it comes to the ultimately unaccountable “flow” of forms and maʿānī from the Universal Intellect to sleeper and interpreter. He undertakes to investigate in a more systematic fashion what causes veridical dreams – and he does so with reference to the normal process of knowledge acquisition.⁶⁵ Given this approach, it seems only natural for Averroes to resort to two other Aristotelian works, Analytica posteriora and De anima, in order to sort out the question that he finds insufficiently treated in Kitāb al-Ḥiss:

[T₅] (1) Let us now consider the productive causes of these two classes of dream-vision.⁶⁶ (2) We shall say: since a true dream-vision indicates knowledge of the existence of something whose existence is naturally unknown to us before we gain this knowledge, and which, at the time when the knowledge is gained, is mostly not [yet] existent, and since this newly gained assent (taṣdīq) belongs to us after we have been ignorant of it, it does not come [to us] as the result of previous knowledge that we had and that could have produced it, nor after a process of thinking and deliberation in the manner in which assenting or conceptual knowledge is [usually] generated, which comes to us as a result of premises. (3) For it has been stated clearly in Analytica posteriora that assenting or conceptual knowledge (taṣdīq/taṣawwur) is naturally preceded by two kinds of knowledge: the productive (fāʿil) and the preparatory (muwaṭṭiʾ). (4) However, this [kind of] knowledge gained during sleep clearly is not preceded by the productive kind [of knowledge]. Whether it is preceded by the preparatory kind will need to be examined.⁶⁷

Averroes here labels the knowledge conveyed by a veridical dream as “assent” (taṣdīq), thus referring to the division of knowledge into the two categories of assenting and conceptual knowledge (taṣdīq/taṣawwur) common in Arabic logic since al-Fārābī.⁶⁸ A veridical dream deals in truths, it tells us that something is (or will be) the case, rather than just helping us with the conceptualisation of things. The reference to this distinction already evokes Analytica posteriora 1.1–2, which is likely to be one of the Greek source texts behind it.⁶⁹ However, Averroes’ explicit reference to Analytica posteriora⁷⁰ (3) aims at another point: Aristotle’s claim that all newly gained knowledge must be based on pre-existing knowledge. Aristotle here furthermore declares that we need “to be already aware of things in two ways,”⁷¹ either by knowing what something is, or by knowing that it exists (or by both). This differentiation is already captured by the distinction between assenting and conceptual knowledge, but Averroes goes one step further and introduces yet another distinction that does not map exactly onto the first one. Both types of knowledge, he finds, are produced from “preparatory” (muwaṭṭiʾ) knowledge and “productive” (fāʿil) knowledge.⁷² Averroes does not explain the distinction any further, but what he seems to have in mind in the present context is the distinction between knowledge of the necessary concepts and the particular, sensible premises on the one hand (i.e., the preparatory knowledge),⁷³ and the decisive piece of knowledge that brings about a new insight on the basis of that preparatory knowledge, thus effecting the move from ignorance and potential knowledge to actual knowledge, on the other.⁷⁴ In the case of true dream-visions foretelling the future, we are dealing with something that could not be produced naturally by our previous knowledge (2). In particular, we do not possess knowledge that could act as a productive cause of such knowledge (4) (or else we could not remain ignorant about it). Where does it come from, then? Averroes concludes that we must acquire it in a manner analogous to the one other case in which we obtain knowledge without possessing the productive knowledge ourselves: the first principles that are fundamental to our intellectual operations, and that are provided by the Agent Intellect:

[T₆] (1) If this knowledge is gained by us after we have been ignorant of it, and is actually present after having been in the state of potentiality, and if there is no knowledge in us that could produce this knowledge, it is clear that our acquisition of this knowledge is like the acquisition of the first premises. (2) If it is so, then it follows necessarily that both have one and the same productive [cause], belonging to one and the same genus. (3) And since it has already been explained in the general accounts⁷⁵ that every thing that emerges from potentiality into actuality has a productive cause that lets it emerge [into actuality] and that must needs be of the same genus as the thing which is emerging from potentiality into actuality, it follows necessarily that the productive [cause] of this knowledge is actual intellect (ʿaql bi-l-fiʿl), and the very same one that provides [us with] the universal principles in theoretical matters, and whose existence has been explained in De anima; (4) for both instances of giving belong to the same genus.⁷⁶

Placing the knowledge acquired in a veridical dream on a par with the knowledge of first principles, just on account of the fact that both types of knowledge are gained in the absence of any previous knowledge that could produce them, turns out to be an ingenious move on Averroes’ part. The argument, developed in (3), leading to the conclusion that the productive cause of such knowledge must be the Agent Intellect, would not that obviously have worked for a type of knowledge that concerns particulars rather than intelligible universals. It is thus only by piggy-backing on the primary principles that true dream-visions can claim their cause to be the Agent Intellect – under the assumption (4) that “both instances of giving belong to the same genus” (a point we will come back to below). The crucial difference is touched upon by Averroes in the continuation of the text, where he states – as if it were an issue of minor importance – that in the case of a divinatory dream, the Intellect furnishes us not with principles for acquiring knowledge, but directly with the specific instance of knowledge itself:

[T₇] The only difference between them is that in the field of theoretical knowledge, it provides the universal first principles that produce the as yet unknown knowledge, whereas in our case here it provides the as yet unknown knowledge [itself], without mediation.⁷⁷

The most problematic aspect of this knowledge (masked here somewhat by Averroes’ formulation), and perhaps the most problematic aspect of the entire theory of divinatory dreaming that springs from Kitāb al-Ḥiss, is, however, its particular nature. The adaptation of the Parva naturalia does not broach the question of how the Universal Intellect is supposed to convey particular forms or maʿānī to dreamer and interpreter; the text does not show any awareness of the problem.⁷⁸ Within the argument in texts T₆ and T₇, Averroes may seem to play down the difficulty; later, however, he addresses it head-on, laying out the difficulty in clear detail:

[T₈] (1) We say: if it appears that what provides this knowledge is an intellect free from matter, and given that it has been explained in the divine sciences that these separate intellects only think universal natures and can only provide things similar to what is within their own substances, it will not be possible for them to provide [us with] an individual maʿnā at all, since the perception of such a particular maʿnā is not in their natures. (2) Those universal forms are only individualised in matter, I mean in the sense that they can only subsist in matter. (3) If the separate intellects possessed individual perception, they would necessarily be material, so that they could only perform their activity through active and passive contact. (4) If, however, those intellects do not think the individual maʿānī, then how, I beg to know, does the Agent Intellect (al-ʿaql al-fāʿil) provide this individual form that is specific to a certain time and location and to one [particular] group of people, or to one individual of that group?⁷⁹

This question is then linked to a second one: why is the knowledge conveyed in a veridical dream only ever granted to someone concerned with the events and things in question, rather than to any or every person, in the way in which knowledge of universals is open to any human being?

[T₉] (1) For we see that a man only perceives such things, and gets warnings in [his] sleep about their occurrence before they happen, in so far as they are specific to his body or soul or to his relations or the people of his city or nation, or generally to what he knows. (2) Thus there are doubts here on two counts: for one thing, how are particular things acquired from the universal nature; and secondly, why are these acts of providing specific particulars specific to the person to whom the knowledge about them is given?⁸⁰

Averroes devotes a lengthy and careful discussion to the first problem, in a long passage that is completely independent of Kitāb al-Ḥiss,⁸¹ but in which he again makes explicit use of other works from the Aristotelian corpus: the Physica and De generatione et corruptione,⁸² as well as what Averroes here calls Kitāb al-Ḥayawān wa-l-nabāt (Book of Animals and Plants), the specific reference probably being to book one of De partibus animalium.⁸³

Briefly put, Averroes explains that the concrete particulars which are predicted by veridical dreams have determinate causes (rather than occurring by chance). Therefore they have a “universal intelligible nature”⁸⁴ which is their primary cause and can be comprehended intellectually. As it happens, we cannot in fact comprehend these things because their origins are too far removed from us in time and we do not know their determinate causes (even though these do exist).⁸⁵

What takes place when a veridical dream occurs, Averroes goes on to argue, is not that the Intellect provides us with the particular form and/or maʿnā of the event concerned (as it is seemingly claimed in Kitāb al-Ḥiss). Instead the Intellect conveys the “universal nature” accounting for the causation of the particular future event; the imaginative faculty of the soul will receive it as a particular, since the faculty is itself enmattered – in an analogous way to the reception of forms (specifically souls) in matter. The process thus resembles very much that of regular knowledge acquisition:⁸⁶

[T₁₀] (1) If all this has been confirmed, it will not be objectionable [to claim] that the separate Intellect provides the imaginative soul with the universal nature which belongs to that individual thing that will come into being, I mean: the intelligible [account] of its causes; then the imaginative soul will receive it as a particular by virtue of its [i.e., the soul’s] being in matter. (2) Sometimes it will receive the individual of that intelligible itself in reality, and sometimes it will receive an imitation of it. (3) Just as [the Intellect] provides the perfections of the soul as universals and matter receives them as particulars, so it gives, in this case, the final perfection to the imaginative faculty as a universal, and the soul receives it as a particular.⁸⁷

With this account of divinatory dreaming, can Averroes maintain what he has said in T₇, that is, that in the case of a veridical dream, the Intellect gives us knowledge about the particular thing or event concerned, without mediation? Strictly speaking, no, if one considers the issue from the perspective of the giver: the Intellect gives a universal (3). However, the matter looks different when regarded from the perspective of the recipient. The imaginative faculty receives the knowledge as a particular, and directly so. It does not need to perform any intermediary steps of deliberation and thinking in order to arrive at the result. In that sense, it is indeed given a specific instance of particular knowledge, without mediation. At the same time, Averroes’ move to say that the Intellect gives a universal helps with the argument in T₆, especially with the important premise that “both instances of giving [i.e., primary premises and divinatory dreams respectively] belong to the same genus” (4).

This solution to the first puzzle raised in T₉ would allow Averroes to hold on to the thesis that divinatory dreams reveal knowledge about particulars, while at the same time avoiding the ascription of knowledge of particulars to a separate intellect. However, his suggestion is not entirely unproblematic. To begin with, there is the question how the truth of the prediction can be guaranteed – usually the task of the divine source of a veridical dream. In Kitāb al-Ḥiss this is achieved by positing a direct relation of representation holding between “intellectual form” and dream as well as between “intellectual form” and future perceptible (T₁), with the unacknowledged difficulty that this seems to require “intellectual forms” of particular character (the very problem Averroes is trying to redress). As Averroes apparently envisages the process, the Agent Intellect provides a universal intelligible that, in combination with knowledge of certain sensible premises, brings about perfect knowledge of the causal chain⁸⁸ leading up to a certain event – manifesting itself in the dream representing the outcome. (Averroes compares this to a physician drawing inferences concerning a patient’s future state of health by applying his universal knowledge of medicine to the particular situation of the patient.)⁸⁹ The necessary sensible, particular premises are, however, not conveyed as part of the dream (and certainly not by the Intellect) but are supplied by the dreamer’s “preparatory knowledge.”

This preparatory knowledge, which was touched upon briefly in T₅, is brought up by Averroes explicitly in relation to the second puzzle raised in T₉, that is, why veridical dreams only occur to people concerned with their subject matter. As it turns out, the preparatory knowledge is the decisive factor in this respect: only those people who have a link to and an interest in the particular event that is being foretold will have the necessary preparatory knowledge which, together with the productive knowledge supplied by the Intellect, will result in actual knowledge about the event.

[T₁₁] Why is it that of [all] these things, a person will only perceive what is specific to his time and place, land, and people, but not all the other particular things that share with them [their] universal nature? The reason for this is that in case of such a perception, one of the two types of knowledge that precede assent must doubtless be present in the person, namely the knowledge that prepares for the assent, I mean the knowledge of prior conceptualisation. For such information and knowledge will only occur to a person as far as it concerns individuals he has known before and in particular those for whom he has already developed a concern. With respect to those [individuals] that are unknown to him, however, knowledge as to what will happen to such an individual cannot possibly come to him. For even though it is not a condition for such an assent that it be preceded by any productive knowledge in the person [concerned], it inevitably is a condition for it that it be preceded by preparatory knowledge.⁹⁰

This contrasts with Kitāb al-Ḥiss, where the fact that a divinatory dream is experienced by a specific individual is seen as the direct result of God’s providential care, the cause behind such dreams, which come about quite independently of any previous knowledge of the dreamer. Moreover, Averroes’ source does not even conceive of such dreams as restricted to people who are directly affected by their content, but is more open as to the scope of their message:

[T₁₂] If someone says: “Why does God let these [forms] appear within dream-vision?,” we say: because their appearance constitutes signs (āyāt) and indications (ʿalāmāt) and an alert for the particular soul. Sometimes such a dream-vision is an alert for a particular soul only; sometimes it is a sign and an indication of some event that will happen in the [whole] world, and sometimes it is [a sign] of something that will happen to a certain man specifically, be it a punishment, or some good that will come to him, or some evil that will befall him. Sometimes the dream-vision concerns both him and the [whole] world together, and sometimes it concerns something else.⁹¹

In restricting dreams to people with the relevant preparatory knowledge, Averroes thus deviates once again from his source. He does so for a reason, though: this step is vital for his explanation of veridical dreaming as analogous to ordinary knowledge acquisition through the interplay of the necessary preparatory knowledge and the universal knowledge provided by the Agent Intellect. His account further contrasts with Kitāb al-Ḥiss in that it goes beyond a mere reliance on the thought that those who will be affected by a future event will benefit most from foreknowledge about it – which would make divine providence the sole decisive factor in the matter. Within Averroes’ attempt to explain veridical dreaming as a case of knowledge acquisition there is not much room for such an element. After all, the mere fact that we may reap benefit from gaining knowledge as such, sadly, plays no causal role in our success in acquiring it. The appropriate previous “preparatory” knowledge, however, obviously does. Averroes thus has identified a factor that can plausibly explain why the universal provided by the Intellect will be converted into a veridical dream in the case of one person, but not the other – and why veridical dreams only ever concern matters that the dreamer is directly involved with (as is his contention).

However, the notion of preparatory knowledge also introduces certain problems. It is not intuitively plausible that the truth of such a divinatory dream could be guaranteed convincingly, if so much work is to be done by the preparatory knowledge of the dreamer – especially considering that divinatory dreaming was supposed to be an ubiquitous phenomenon, experienced by all and sundry rather than by particularly well-informed people or trained experts (like the physician featuring in Averroes’ comparison).⁹²

Another worry in this context concerns once again the claim that in a veridical dream the Agent Intellect conveys the particular knowledge directly, “without mediation” (T₇). Perhaps Averroes merely refers to the fact that there is no conscious thinking process involved; nevertheless it is hard not to see the final outcome, the knowledge conveyed by the dream, as somehow or other mediated by the preparatory knowledge of the dreamer.

This point touches on the question of what exactly must happen in order to bring such a dream about and what renders it the miraculous phenomenon that it is supposed to be.⁹³ It is the Agent Intellect that is portrayed as the source of the revelation, the giver of the dream. However, considering that its contribution is simply an intelligible universal, this cannot take the form of a separate, specific act of conveying that intelligible to the dreamer; in principle, this universal will be constantly and universally available. This seems to give the preparatory knowledge an even more crucial role. Do we have to assume a certain automatism, with veridical dreams ensuing once people have acquired the suitable preparatory knowledge? Not quite; as Averroes points out, the ability to have true dreams also depends on a person’s imaginative faculty, which again is dependent on his or her humoural constitution;⁹⁴ a further factor is the specific situation of the imaginative faculty during sleep.⁹⁵ But in any case it seems that no further divine act is required and, moreover, that there is no specific act solely responsible for the occurrence of a particular divinatory dream, which instead appears to be the outcome of the interplay of several factors. While this may well be regarded as a manifestation of the general workings of divine providence in the world,⁹⁶ it would still constitute a quite radical revocation of the idea, maintained in Kitāb al-Ḥiss, that such dreams are individually intended manifestations of God’s solicitude for mankind. (If anything, such solicitude must manifest itself more in ensuring that a person assembles the necessary preparatory knowledge than in the supply of the anyway generally available intelligible universal through the Intellect.) However, this may well have been a price that Averroes was very willing to pay.⁹⁷

With his investigation into the causes of divinatory dreams Averroes therefore develops an interpretation of the theory found in Kitāb al-Ḥiss that alters it considerably. Discussing veridical dreams as cases of knowledge acquisition and grounding them in Aristotelian epistemology allows him not only to clarify (and rectify) a number of points that are problematic or unsatisfactory in his source text, but also to reduce the aura of the sublime, miraculous, and unaccountable that at times surrounds the topic in Kitāb al-Ḥiss. Veridical dreaming becomes, at least in principle, fully explicable within Averroes’ Aristotelian framework, without leaving the need to appeal to a deity that would determine the receivers of such dreams on an individual basis, or having to leave open the question of how exactly the individual instances of veridical dreams could be caused if a providential force operating in nature were solely responsible for them. As with knowledge acquisition in general, veridical dreams are a result of the activity of the Agent Intellect, and of our own individual aptitude, preparedness, and situation. What gives this kind of knowledge acquisition its miraculous air is that it circumvents the normal processes of deliberation and learning (which makes it available to people of all shades of intellectual aptitude), resulting in a dream image that may well be enigmatic rather than in propositional knowledge, and taking place at a time when intellectual operations are shut down.⁹⁸ However, it is miraculous in the sense of being unusual rather than being inexplicable;⁹⁹ Averroes at least attempts to account for every aspect and detail within his epistemological and psychological framework. In this, he shows himself to be very true to the Aristotelian spirit of De divinatione per somnum, and rather detached from that of Kitāb al-Ḥiss. One may hence wonder how much of the “divine” veridical dreams really retain in his view. However, while the general sentiment of his treatment is pretty close to Aristotle’s – veridical dreams ought to be explained within the rational limits of philosophy – there is a difference that may be crucial here. Aristotle attempts to explain veridical dreams within the framework of perception and imagination, whereas Averroes does so within the framework of knowledge acquisition. He thus takes much more seriously the claim to an independently guaranteed truth that attaches to these dreams; and given that in his world, knowledge and truth, and the intellectual source of the dream belong to the realm of the divine, the label “divinatory” may yet be merited by veridical dreams within his system of thought.

5 Veridical Dreams as Imitations of Reality and Their Miraculous Character

Divinatory dreams, especially if considered as vehicles of knowledge acquisition, have an inconvenient feature: they are not always easily understood (hence the need for dream interpreters). This point has been touched upon already in T₁₀ (2), and leads to the question in which way exactly a true dream “indicates knowledge.” Put differently: why does a veridical dream sometimes show us things not as they really “look like,” but in a different form? The problem reflects the circumstance that people did not expect veridical dreams to depict future events exactly, but understood them as symbolic dreams that needed to be interpreted.¹⁰⁰

The question of why divinatory dreams do not depict things exactly as they will present themselves in real life, once they come to pass, is already broached by Kitāb al-Ḥiss:

[T_13a] Now if someone says: “[Assuming] a man sees the land of the Franks, the land of the Greeks, or Africa in a dream, without ever having seen them with [his] sense[s]. Then perhaps when he later does see them with [his] sense[s], they will not be as he has seen them in his dream. In that case, one of two alternative explanations must apply: either the form he has seen in his dream is not like the one he has seen during waking; or it is the same, but the formative [faculty] has committed a mistake” […].¹⁰¹

Averroes puts it more dryly:

[T_13b] Why, in most cases, does the imaginative faculty not convey the real individual maʿnā which falls under the universal which the intellect emanates, but only conveys the maʿnā that imitates it?¹⁰²

Kitāb al-Ḥiss suggests as a solution that the three elements concerned – “intellectual form,” “spiritual form” (dream), and “corporeal form” (actual event) – stand in a relation of representation to each other, being images of each other:

[T_14a] […] we will answer: (1) The formative [faculty] has not been mistaken about the form of this city which he has seen within the confines of dream-vision, <as> it is indeed the one which he has [later] seen during waking, (2) because every corporeal object of perception has two forms, one spiritual and one corporeal, the spiritual form being inside the corporeal form. Just as the corporeal form of the city is an image of the spiritual form of the city, which is inside it, likewise the spiritual form is an image of the intellectual form. Hence, if you see a form within the confines of dream-vision in the way I have described then you will only see your internal form, because the intellect has dressed up its own [i.e., the intellectual] form, and has embellished it with spiritual words, whereupon the common sense conveys those words to the formative [faculty] so that it represents that form. Hence when the common sense sees those words and recognises them as it here recognises corporeal words and their written representation (rasm), it presents them to [the faculty of] memory.¹⁰³

Remarkably, this approach places the spiritual form “inside the corporeal form,” while regarding the corporeal form as an image of the spiritual form (which again is an image of the intellectual form) (1). This has to do with the adaptor’s spiritual-corporeal hierarchy: the corporeal form is equated with “shells” that have to be cleared away to arrive at the pure reality of the spiritual “core.” The existence of these two forms explains why the representation one sees in a dream may not look identical to the “corporeal” manifestation of the thing in question. What is particularly interesting for us in this context is the rather Platonic concept of imitation: the corporeal is the image of the spiritual, which means: the “real” event in this world will be the imitation of the dream (or at the very least of the corresponding “intellectual form”) rather than the other way round.¹⁰⁴

Averroes reinterprets the two-form-solution offered by Kitāb al-Ḥiss, dropping the Platonism in the process:

[T_14b] This is because the perceptible object has two forms: a spiritual one, which is the form that imitates it; and a corporeal one, which is the form of the perceptible thing itself, not the form that imitates it. The form that imitates [it] is more spiritual, because it is closer to the universal nature than the real form of the thing. Therefore the imaginative soul receives the intelligible maʿnā with the most perfect spirituality its substance is capable of receiving. Sometimes it may receive it in a corporeal manner, so that the dreamer sees the form itself in his sleep, not what imitates it.¹⁰⁵

Thus, he emphasises explicitly and repeatedly that the spiritual form (i.e., the dream) is the image, whereas the corporeal form is the reality, the “real form of the thing.” The imitating form¹⁰⁶ is “spiritual” because it is closer to the universal; this proximity, however, does not mean that it is “imitated” by the corporeal form. Averroes’ additional remark that sometimes the imagination does receive the corporeal form itself (i.e., the dream depicts things just as they will look like in real life) leaves it open to speculation whether a greater degree of spirituality is always a better (or at least more useful) thing – after all, veridical dreaming is all about gaining information about one’s real future life, and from this point of view an imitation may well seem only second-best to a true depiction. Thus, the term “spiritual” again appears much less elevated in Averroes than in the text he is paraphrasing.

Averroes still acknowledges the special achievement of the imaginative faculty during sleep, linking it to spirituality: “the action of the imaginative faculty is more perfect and more spiritual during sleep.”¹⁰⁷ However, he does not seek the reason for this perfection in the loftiness of the faculty’s objects (i.e., veridical dreams), but merely in the fact that during sleep the external senses are at rest, which allows the imaginative faculty to grow stronger because “the soul” can now concentrate on it.¹⁰⁸ Thus, the perfection achieved by imagination during sleep appears to be grounded solely in the fact that the imaginative faculty can now carry out its proper function uninhibited.

The knowledge of the future a veridical dream can offer is miraculous because it fills a gap at the point where our normal mental faculties, in particular the rational faculty, reach their limits. It is also particularly beneficial because it allows the person concerned to prepare for things to come. It is, Averroes says, due to providence.¹⁰⁹ While this is not far from the contention expressed in Kitāb al-Ḥiss that God, with His providential care for mankind, is the ultimate cause behind veridical dreams,¹¹⁰ Averroes’ invocation of an abstract principle of providence instead of a direct reference to God nevertheless bespeaks a rather reserved attitude towards the traditional religious ways of accounting for such dreams.¹¹¹

Nevertheless Averroes does not hesitate to embed veridical dreams within a religious context; he mentions, for example, the story of Joseph interpreting Pharaoh’s dream from Sura 12 of the Quran,¹¹² and a tradition concerning a vision seen by the caliph ʿUmar Ibn al-Khaṭṭāb.¹¹³ However, where Averroes does mention God as possibly involved in veridical dreaming, he does so in a rather guarded manner, as if wanting to dissociate himself from the suggestion: “people believe that dream-visions come from the angels, divination from the jinn, and revelation from God”;¹¹⁴ “[prophecy] has been related to the Deity and to the divine beings, that is, the angels.”¹¹⁵ In any case it is the angels that are here named as causes of prophetic dreams. Since angels can be read as standing for the cosmic intellects, this would not take Averroes far away from his philosophical account of divination caused by the Agent Intellect.

6 Dream Interpretation and the Interpreter

When it comes to the question of dream interpretation and the figure of the dream interpreter, we can again observe a significant difference between Averroes’ account and Kitāb al-Ḥiss wa-l-maḥsūs. According to Kitāb al-Ḥiss, the Intellect conveys to the interpreter the maʿnā of the dream that it conveys to the dreamer; that is, the interpreter will know about the future perceptible event in question without being given an image-like representation of it. In addition, he will also be able to relate dream and maʿnā correctly to each other. It is because dream and maʿnā as well as the corporeal form of the future perceptible share the relation to the intellect that the interpretation will be correct.¹¹⁶ Like the dreamer, the interpreter is entirely dependent on the revelation from the Intellect:

[T₁₅] Sometimes the Intellect flows upon the interpreter with [those] spiritual words, then his tongue will pronounce them, while he will see that he is the one who [correctly] interprets that dream-vision. At other times the Intellect does not flow upon the interpreter; then he will commit mistakes and will not know what to say or what he should interpret.¹¹⁷

In accordance with its focus on the corporeal-spiritual divide, Kitāb al-Ḥiss describes the interpreter and his qualifying features first and foremost in terms of “spirituality”:

[T₁₆] The Intellect only flows upon the interpreter for one of two reasons: either the interpreter is spiritual, then the Intellect flows upon him because of his spirituality. Or it flows upon him because of signs (āyāt) that [shall?] become manifest in the world.¹¹⁸

Kitāb al-Ḥiss does not expand on the exact meaning of “spirituality” here. Again the vacillating term allows the adaptor to bring several associations into play:¹¹⁹ it suggests an affinity with (and hence competence for) the “spiritual” forms and maʿānī, perhaps by virtue of having particularly well developed “spiritual faculties,” or good quality “spirit”; it intimates concern for, and familiarity with “spiritual,” that is, incorporeal (intellectual and divine) things. However, this close association with the realm of the spiritual and divine also has a distinct ethical ring, implying a disregard for, or even the renunciation of, worldly, corporeal things in favour of striving for a life untainted by sin. This idea is underpinned by further remarks indicating that the interpreter is supposed to “purify and refine his soul” and “free his body from dirt and impurities.”¹²⁰

The second possible explanation presented for the Intellect’s “flowing” unto the interpreter is somewhat more obscure. It could mean that regardless of the interpreter’s preparedness in terms of spirituality, the interpretation of a dream could be revealed in order to secure fulfilment of the providential purpose of divinatory dreams, that is, the transmission of information about particular future events. In this case it would point once more to God’s providential care that lies behind the entire phenomenon in the first place.¹²¹ Alternatively, it could suggest another kind of preparedness on the part of the interpreter: the ability to read the dream in the light of other signs that he has observed in the world. While a plausible reading of the sentence in itself, it is not particularly convincing in the context of Kitāb al-Ḥiss, where this point is not picked up anywhere else, and where there is little indication that observations about this world could make any positive contribution in this matter.

Whether correct or not, this would probably have been a reading more to the liking of Averroes, who provides a slightly different account of the interpreter’s qualifications:

[T₁₇] (1) The interpreter is a man whose soul is naturally disposed to understand the imitations that occur in a dream-vision; (2) he is the person on whom the intellect emanates the corporeal maʿānī which in sleep are imitated by spiritual maʿānī. (3) Among the conditions applying to him is that he be knowledgeable about the [dream-]imitations that are common to all nations, and about the imitations that are specific to each nation and to each class of people […]. (4) Also, as Aristotle says, it behoves the interpreter to put his soul in proper condition through thinking and theoretical reflection, and his body through cleanliness, and to be chaste and not to incline to traits of the animal soul, and to be spiritual.¹²²

Rather than merely invoking “spirituality,” Averroes thinks that successful dream interpretation requires first of all a natural aptitude to understand dream images as imitations (1). However, talent is not everything: the interpreter must also know what the various types of dream images stand for (3). As Averroes describes this knowledge, it seems to be something one has to learn rather than understanding it intuitively, or receiving it through revelation. With an apparent nod to the oneirocritical tradition which seeks to classify dream images as symbols for various types of events or things, Averroes explains that what dream images signify (i.e., imitate) may vary from nation to nation or from one social group to the other. This is partly due to natural characteristics obtaining in the mental faculties of such groups as well as in their environment, but also to differences in culture and traditions.¹²³ Nevertheless such preparedness on the part of the interpreter is only a necessary rather than a sufficient condition (3) for his ability to interpret a particular dream, given that the Agent Intellect is crucially involved in Averroes’ account too: it provides the interpreter with the “corporeal maʿnā” of which the dreamer has seen an imitation, a “spiritual maʿnā” (or form) (2).

The expression “corporeal maʿnā” must strike one as odd, especially given the fact that in Kitāb al-Ḥiss maʿnā is the most spiritual entity the post-sensatory faculties deal with and is said to retain no trace of corporeality. Nevertheless, it is sufficiently clear what Averroes must mean: the interpreter can identify the “corporeal” thing or event in the real world that the dream imitates through a “spiritual” form/maʿnā. Obviously, there will be no “corporeal” entity in the mind of the dreamer. Averroes uses the term “corporeal” here as a marker for reality as opposed to imitation, which is “spiritual.” This again confirms that Averroes does not follow Kitāb al-Ḥiss in its Platonic tendency to suppose that the veridical dream gives us access to a higher, more spiritual, and more “real” reality of which the corporeal world is an image, a “shell.” The reality that is crucial here for him is our “corporeal” reality.

The emphasis on the interpreter’s skills in reading imitations furthermore suggests that even after having been given knowledge of the unknown future perceptible event there still remains a task for the interpreter that requires him to tap into his preparatory knowledge: assigning the right meaning to the right dream. And while Averroes does cite “Aristotle” as saying that the interpreter needs to attend to cleanliness, chastity, and spirituality, he also mentions “thinking and reflection” as commendable habits (4), something we do not find as such in Kitāb al-Ḥiss in this context.

Again we see that Averroes interprets the phenomenon of veridical dreaming as a case of acquiring knowledge, not in the manner of a more or less passive reception of a revelation that transcends natural boundaries, but in a manner closely analogous and directly related to our usual way of gaining knowledge about our world. He furthermore does not take over the Platonic order of imitation and reality, but considers the “corporeal” perceptible, the thing or event in this world, as the relevant “reality” about which divinatory dreams help us to acquire knowledge and which may be imitated by the images of a veridical dream.

7 Concluding Remarks

Aristotle’s sceptical position on veridical dreams was an unusual one to take in Greek antiquity as well as in the mediaeval Islamic world – not just within society at large, but also among philosophers. It would have been intriguing to see what Averroes, not one to shy away from defending controversial philosophical tenets, would have made of Aristotle’s true stance on divinatory dreaming. Alas, thanks to the adaptor of Kitāb al-Ḥiss we will never know. Nor will it be possible to determine with certitude what exactly Averroes thought of Kitāb al-Ḥiss. That he did not simply dismiss it, or its attribution to Aristotle,¹²⁴ is evident not just from his Explanatory Paraphrase, but also from the role the theory of the post-sensatory faculties – which is associated with Kitāb al-Ḥiss – plays within his psychology more generally. A detailed investigation of the exact shape in which the ideas contained in Kitāb al-Ḥiss enter, for instance, his Long Commentary on De anima may further contribute to our understanding of his attitude towards the text. Here I hope to have shown that Averroes seems to exhibit at least a degree of scepticism as to the adaptation’s contents – if not concerning the main lines of its theory, then at least with respect to some of its details. Confronted with the Neoplatonising account of divinatory dreaming put forward in Kitāb al-Ḥiss, he does not address the problematic points openly, but silently uses his knowledge of Aristotle’s works and of the interpretative efforts of his predecessors as well as his philosophical judgement to reinterpret his source in a way that allows him to preserve important Aristotelian tenets and to develop an outlook on the topic that is more in line with Aristotle’s philosophy than the Arabic text transmitted under the Stagirite’s name.