3.1 Why Dreams Are Taken to Be Real

The problem of how we may prove that we are not dreaming right now is mentioned briefly by Aristotle in the Metaphysics¹⁴ and is described as an artificial problem that misleadingly suggests the need of a demonstration in order to be resolved. Aristotle argues that to ask for proof in order to distinguish between one’s dreaming and waking is to have things backwards, since the fundamental distinction between veracity and error is founded on circumstances that do not need proof. The problem of dream-scepticism is discussed in the context of Protagorean relativism and echoes Plato’s formulation of the problem in Theaetetus 157e–58e. Aristotle holds that there are normal circumstances of waking life that serve as paradigmatic cases of authoritative awareness in relation to more or less distorted apprehensions of reality due to a variety of causes. Aristotle seems to endorse the view that under normal circumstances we can recognise that we are awake and that the state of waking is more authoritative than sleep or dreaming. However, this does not rule out that we occasionally may be deceived by dreams during sleep.

To reduce overlaps with Gregoric’s contribution in this volume I shall highlight a set of themes that are not extensively treated in his paper, including the nature of delusional dreaming as a forced mode of deception and varieties of deflationary notions of trust. It will be shown that the concept of delusional dreaming, in one interpretation, illuminates Aristotle’s account of why dreams are cognised as real. Here is what Aristotle has to say about why people tend to mistake dream-images (phantásmata) for real events:

Each of these [phantasms], as has been said, is a remnant of the actual sense-impression, and is still present within, even when the real one [viz. sense-impression] has departed. Thus, it is true to say that it is like Coriscus, even though it is not Coriscus. While one was perceiving, one’s ruling and judging part was saying not that the sense-impression is Coriscus, but because of that impression, that the actual person out there is Coriscus. The part that says this while it is actually perceiving (unless it is completely held in check (katéchetai) by the blood) is moved by the movements in the sense-organs, as if it were perceiving.¹⁵

In short, dream-appearances derive from sense-impressions and linger on and become apparent in sleep because ordinary perception is shut down. One central idea is that, like veridical perceptual states, dream-appearances present information about how things appear to be in the world, even if false. For example, when we perceive Coriscus the senses display (the real) Coriscus as-if-present. Likewise, the dream (of Coriscus) involves an appearance of the real Coriscus even if the real Coriscus no longer causes the presence of the sense-impression that persists in sleep.

Aristotle attributes a primitive form of judgement (dokeîn) to the senses, roughly the senses have the capacity to assess how things are in the world, which should not be conflated with proper belief (dóxa).¹⁶ Sight is said to be more authoritative than touch because if touch stood alone, one single object would be taken as two when the object is sensed by crossed fingers.¹⁷ Furthermore, appearances in perception sometimes contradict what is believed or known about the world. For example, the sun appears to be one foot across. In such cases there are conflicting cognitive assessments.¹⁸ Thus, appearances may be challenged by many different kinds of cognitive assessments where some assessments are more trustworthy than others. For example, vision normally outranks touch, and accumulated knowledge about the world is superior to how things superficially look. It seems plausible to assume that what Aristotle refers to as superior cognition in terms of the “the ruler” (tò kýrion) may vary with the circumstances. For example, vision loses its status as the authoritative sense in a pitch-black environment. Exactly how Aristotle’s discussion of hierarchical cognitive assessments relate to the conditions of deception in dreaming is not fully transparent. I shall discuss this problem briefly below, especially in connection with Radulphus Brito’s (d. 1320/21) account.

Aristotle seems to mean that the highest instance of judgement, in relation to normal observational circumstances, is sufficient for reliable judgements. If the cognitive machinery works the way it is supposed to (that is, the subject is not diseased, sleeping, intoxicated, and so forth), the subject would know when he is awake, and accordingly he would know that he is not asleep and dreaming. The point is not that there are no ambiguous cases between sleep and waking, but rather that, given the radical difference between the cognitive conditions of sleep and wakefulness in their paradigmatic manifestations, it makes good sense to suppose that we can be aware of being awake when we in fact are awake, that is, in a state where the mind can exercise the full range of its cognitive powers.

Conversely, Aristotle’s example of deception regarding the reality of the dream seems to rely on a similar assumption that all other cognitive functions are shut down except the awareness of the dream-appearance. In the case of humans, the intellect as well as lower capacities are inactive. Now, it seems reasonable to assume that animals can be deceived about the reality of dreams even if they lack rational capacities. This point is important because it shows that the conditions for deception in dreaming do not include having an intellect, even if the operation of the intellect may be of help in judgements that experiences are illusory. For instance, (1) a dog may be deceived about the reality of the dream, that is, not be aware of that the dream involves an illusory world, (2) when the dog wakes up it becomes aware of the real world, but this does not necessarily imply (3) that the dog has the capacity to be aware of a dream (or any other experience) as an illusory entity. In other words, deception regarding the reality of the dream does not presuppose the cognitive powers to know or suspect that we are sometimes deceived.

Aristotle’s examples of dream deception in which there is no opposing cognition in relation to the appearing phántasma probably reflects a typical condition in which deception occurs. We do not receive any information about what is minimally sufficient for snapping out of the deceptional state in which dreams are taken to be real. Nevertheless, Aristotle is quite clear that the capacity for belief-formation is restrained in the state of sleep. On one interpretation, this means that the dream is neither believed to be real nor believed to be unreal (this is the non-doxastic interpretation of deception in dreaming):¹⁹

And does judgement sometimes declare it an illusion, as it does for waking people, while at other times it is held in check (katéchetai) and follows along with the appearances (phantásmata)?²⁰

And in a related passage:

For in general the starting-point affirms the report from each sense, provided that some other, more authoritative one does not contradict it. In every case, then, something appears, yet what appears is not in every case judged to be real; it is, though, if the critical part is held in check (katéchetai) or fails to move with its own proper movement.²¹

In this second passage Aristotle does not explicitly refer to belief-formation as restrained but only to the “critical part.” So, does he refer to the capacity to form critical beliefs or to any assessment that may oppose a present phántasma? It seems plausible to interpret Aristotle as referring to any opposing cognitive assessments as absent in the typical case of dream-deception. The reference to dóxa as a product of the operating intellect in the first passage, however, is likely to highlight the point of there being a deceptive non-doxastic form of naïve trust in dreams.

As for the question of belief, I follow Gregoric’s view that deception does not require a proper belief. It is a form of unreflecting trust, which involves a tacit uncritical affirmation of whatever appears to be the case. Aristotle maintains that there is an element of low-level judgement (dokeîn) that is conveyed by the senses. The dokeîn-element in sense-perception basically means that the senses, in one very restricted sense, can be said to judge how things are in the external world. This form of low-level judgement should not be conflated with fully developed belief (dóxa) yet it seems sufficient to explain the apprehension of a dream as something real. Now, if the sensory remnants in sleep convey some sort of judgements about how things are, which are manifest as appearances, the dreamer will respond to such appearances as true given the absence of opposing cognitive assessments. This is the non-doxastic interpretation of why the dreamer is deceived regarding the reality of the dream.

However, there is another option. It might be argued that the state of sleep only inhibits the formation of critical beliefs, not the formation of affirmatory beliefs. So when the dreamer responds to the apparently real dream as actually real there is a proper belief in the dream as real. Yet it is difficult to see why the state of sleep would restrain critical belief formation but leave the capacity for affirmatory belief intact. If there can be only confirmatory beliefs there is not much left of the normal belief-generating procedure that is the result of rational evaluation, and consequently it becomes odd to characterise such non-rational assents as proper beliefs.²²

3.2 Realism, Deception, and Dream Bizarreness

As we have seen, Aristotle’s explanation of delusional dreaming involves examples that faithfully mimic sensory qualities and also mundane events, for instance, when it seems as if a well-known person (Coriscus) is present. Aristotle does not explicitly discuss incredible dream-content or cases in which dreamers respond to fantastic dream-content as real, nor does he discuss the cognitive capacity needed in order to be able to assess what is considered as normal, credible, or hard to believe. In other words, Aristotle does not explicitly address the question of why absurd dream-content is accepted without further notice. Nonetheless, it seems reasonable to assume that the dreamer responds to the phántasma as nothing strange because it provides the only present cognitive assessment about how things are. In such simple-minded cognitive states, nothing can be analysed or examined.²³

However, Aristotle discusses aspects of dream-bizarreness from another angle in De insomniis. Strange dream-contents are understood as distortions that are caused by physiological disturbances and may involve deformed, fragmentary, or rearranged sense-impressions to the extent that they no longer resemble the perceived objects they derive from.²⁴

When Aristotle refers to distortions, or grotesque or incoherent dreams, he seems to be focused on strange dream-images in the form of objects, not on fantastic dream-narratives. Two related points can be made in connection with this observation. First, Aristotle’s conception of distorted dreams, just like his conception of undistorted dreams, is object-centred, rather than event-centred (unlike our contemporary notion of dreaming which involves a succession of events, sometimes in the form of stories).²⁵ There is no remark on bizarre sequences of events in dreams in Aristotle’s account.

Further, Aristotle does not make any distinction between (1) what qualitatively appears as real, and (2) what can be reasonably taken to be a realistic experience of an object or an event. For example, a dream of a chimera may appear as real (in a qualitative sensory sense) but may seem unrealistic given what I know about the world. For Aristotle, a dream about a chimera is a distortion, most likely a unity composed of several sense-impressions with different origins.²⁶ A further implication is that the dreamer tends to apprehend a dream-image of a chimera as real, regardless of whether such a creature exists or not. Yet the parts of a chimera-dream correspond to various fragments that derive from mundane sense-impressions. In sum, Aristotle might explain the dream-subject’s tolerance of dream-bizarreness by way of the restrained powers of the intellect or some other inhibited cognitive function that is deactivated by the state of sleep. Even if Aristotle does not explicitly discuss what capacity would be sufficient to identify a dream as odd, incredible, or bizarre, the suggestion that the dream-appearance, as such, involves the only present cognitive assessment, indirectly explains why bizarre elements are ignored or taken as equally real as mundane events.

3.3 The Proper Sense of ‘Unreflective Trust’

Non-doxastic accounts tend as a rule to sound slightly more intellectualistic (that is, doxastic) than they are intended to be. The problem is mainly terminological and concerns the element of assent that may appear to be some kind of top-down attitude taking sense-impressions as its object. The dokeîn-element of a sense impression mentioned above is not something that is added to the sense-impression, but ought to be regarded as an integrated part of it. Hence when the dreamer blindly or non-committedly follows his dream-appearances, the element of trust, conviction, or assent is already there in the very awareness of the appearance itself, provided that there are no other assessments.

Therefore, expressions that aim to capture the non-doxastic interpretation of delusional dreaming conceptualised as an “unreflective acceptance,” “unreflective assent,” “tacit acceptance,” “taking the appearance of realism for granted,” “passive trust,” etc., should not be understood as something added to what is included in the bare awareness of the appearance. Hereafter I shall refer to this non-doxastic type of assent as “unreflective trust.”

We need to make some further qualifications of the relevant Aristotelian sense of ‘unreflective trust’ that is characteristic of delusional dreaming. First, the relevant kind of ‘unreflective trust’ only partly corresponds to similar phenomena we are familiar with from waking life. For example, we usually presuppose (take for granted, or accept) without any critical reflection that the ground under our feet is solid as we walk, if it appears that way. It is arguable that the dreamer tacitly believes, (thinks, presupposes, or takes for granted) that what is experienced during dreaming is real, in the same sense a wakeful person ‘believes’ that the ground under his feet is solid if there are no particular reasons to be uncertain about it. In such cases it becomes reasonable to speak of an absence of doubt rather than the presence of an affirmative belief as something actively produced in response to an appearance. However, it is important to realise that the condition of delusional dreaming does not include any readiness to form critical attitudes as we may do when we tacitly trust that the ground is solid under our feet.

Even naïve or spontaneous beliefs in the existence of the external world are misleading analogues in this context. Why? There is a compelling element in delusional dreaming that distinguishes it from other kinds of unreflecting trust in waking. For instance, when the person is fully awake, he can withdraw, modify, or suspend his trust in bare experience, as he sees fit. For example, if I fall into a camouflaged hole in the ground, I might be very suspicious the next time I stroll around in similar surroundings. Or, even if I do not for a second commit to the belief that there is no external world, I can doubt its existence as part of a sceptical exercise, perhaps to the point where I feel the world to be unreal (without really being deceived by the feeling of unreality). The delusional dreamer, by contrast, cannot withdraw his trust in experience at will for any reason, not even by means of a suspension of judgement or a speculative hypothesis that the present experience is illusory. In sum, the relevant kind of unreflective trust that is characteristic of delusional dreaming, in Aristotle’s account, manifests itself as a rigid, narrow awareness that takes experiences at face value and cannot be modified or altered at will as long as the cognitive inhibitions that are characteristic of the delusional state prevails. Give and take some details, this characterisation seems a plausible option in contemporary theorising.

6.1 Descartes: Deception in Sleep and Waking and the Question of Dream-Belief

In the early modern period Descartes reintroduced the ancient theme of dream-scepticism – widely known as the argument from dreaming.³⁵ Descartes stresses the illusory nature of dreams and raises the problem of how we can know that what appears to be an ordinary experience in waking is not really a dream. The argument from dreaming, in various formulations, is discussed by Descartes mainly as an epistemological problem involving the quest for criteria that guarantee certain knowledge. Nonetheless, one strand in Descartes’ discussion concerns the deceptive nature of dreams. In a well-known passage in the First Meditation Descartes suggests that dreams are unreliable because they occasionally mimic plain experiences in waking. However, it might be useful to highlight some divergences between Descartes’ dream-scepticism and Aristotle’s analysis of delusional dreaming. First, Descartes does not pay any special attention to the fact that dreaming and waking exemplify two radically different states of cognition – at least not explicitly. For instance, Descartes does not seem to endorse the view that we are likely to be deceived about the true nature of our experiences in sleep but not in waking, which is one of Aristotle’s basic assumptions. Second, a tacit supposition in Descartes’ discussion of dream-scepticism is that the capacity for belief-formation is intact in the dream-state. This assumption, as we have seen, is explicitly rejected by Aristotle. Descartes’ and Aristotle’s opposing views on dream-belief will be an important backdrop to more recent discussions on whether genuine belief-formation is possible in the state of sleep – discussions which otherwise may appear more or less obscure.

Let us now turn to an early modern account that more directly resembles Aristotle’s idea that sensory experience is taken as real in the absence of information to the contrary.

6.2 Spinoza: Revisiting Aristotle

In a passage in the Ethics, Spinoza highlights a kind of unreflecting trust that is characteristic for the dream-state – a notion that strongly reflects Aristotle’s conception of a non-doxastic form of trust as described in De insomniis:

So this may be clearly understood, let us conceive a boy imagining a winged horse, and perceiving nothing else. Since this imagination involves the existence of the horse (by Prop. 17, Coroll., Part 2) and the body perceives nothing that takes away the existence of the horse, he will necessarily regard the horse as present, nor will he be able to doubt of its existence, even though he is not certain of it. We experience this daily in dreams, and I do not believe that there is anyone who thinks that, whilst he is dreaming, he has a free power of suspending his judgement about that of which he dreams, and of bringing it about that he does not dream what he dreams he sees. Nevertheless, it happens that even in dreams we suspend judgement, namely when we dream that we are dreaming. Further, I grant that nobody is deceived in so far as he perceives; that is, I grant that the imaginations of the mind, considered in themselves, involve no error (see Prop. 17, Schol., Part 2). But I deny that a man affirms nothing in so far as he perceives. For what is it to perceive a winged horse, other than to affirm wings of a horse? For if the mind were to perceive nothing other than a winged horse, it would regard the horse as present to it, and would have no cause of doubt about its existence and no faculty of dissent, unless its imagination of the winged horse were joined to an idea which takes away the existence of that horse, or because it perceives that the idea of the winged horse that it has is inadequate. Then it will either necessarily negate the existence of the horse, or it will necessarily doubt it.³⁶

Thus, if the mind perceives nothing but a winged horse, then there is nothing that may provide a ground for doubt or disbelief, and so the presence or existence of the horse will be taken for granted. Supposing that the mind perceives nothing but a winged horse, a kind of affirmation of the perceived object’s existence will follow from the mere awareness of it.

Spinoza’s view involves some ideas that resemble Aristotle’s account of delusional dreaming: for instance, (1) there is a compelling type of awareness that includes (a) an apprehension of the imagined horse as real, (b) the inability to doubt its existence or suspend judgement,³⁷ (c) even if there is no certainty of what is apprehended (“certainty” here refers to something that is the result of intellectual assessment). Thus, perceiving (understood to include imagination) considered in isolation from other cognitive features conveys an element of trust in what is perceived. (2) Doubt and disbelief require grounds, and since the awareness of the winged horse is the sole cognitive assessment, without any competing evaluation, the presence of a winged horse is taken as real. A point concerning the example including a winged horse that easily goes unnoticed is that even when such incredible things as winged horses are apprehended, they are apprehended as real. We may also note that Spinoza’s example reflects Radulphus Brito’s discussion of the cognitive conditions that make the dreamer inclined to experience dreams as real even when dreams are completely incoherent or absurd (cf. section three).

6.3 Late Modern Adaptations of the Principle of Unreflecting Trust

A version of Spinoza’s account appears in William James’ Principles of Psychology. James elaborates on the idea of unreflecting trust as a result of absent contradictory information, and his immediate source appears to be Spinoza’s Ethics. He discusses the idea in very general terms – not specifically in the context of delusional dreaming – and makes no explicit reference to Aristotle. James gives a rather general characterisation of belief in terms of a mental state that involves some degree of assuredness, certainty, or conviction.³⁸ Belief, or a sense of reality is a feeling close to an emotion of conviction.³⁹ James maintains that the true psychological opposites of belief are doubt and inquiry, not disbelief,⁴⁰ and he rhetorically asks what it would mean if a simple apprehension of something were to be considered as not real:

Suppose a new-born mind, entirely blank and waiting for experience to begin. Suppose that it begins in the form of a visual impression (whether faint or vivid is immaterial) of a lighted candle against a dark background, and nothing else, so that whilst this image lasts it constitutes the entire universe known to the mind in question. Suppose moreover (to simplify the hypothesis), that the candle is only imagery, and that no “original” of it is recognized by us psychologists outside. Will this hallucinatory candle be believed in, will it have a real existence for the mind?

What possible sense (for that mind) would a suspicion have that the candle was not real? What would doubt or disbelief of it imply?⁴¹

James concludes:

The sense that anything we think of is unreal can only come, then, when that thing is contradicted by some other thing of which we think. Any object which remains uncontradicted is ipso facto believed and posited as absolute reality.⁴²

The most basic forms of belief occur when there is no contradictory information that opposes the existence of the object of awareness. Such primitive beliefs in the reality or existence of something do not involve some reference to some special property or quality that makes it real. Rather, any object of awareness, in the absence of contradictory information, will be apprehended as real. The idea here is that primitive unreflecting trust does not imply fully developed belief, while doubt and disbelief imply a fully operational ability to form beliefs.

James considers a case in which we apprehend something in an epistemically detached way and on the basis of rational evaluation make an affirmative or critical judgement:

The having and the crediting of an idea do not always coalesce; for often we first suppose and then believe; first play with the notion, frame the hypothesis, and then affirm the existence, of an object of thought. And we are quite conscious of the succession of the two mental acts. But these cases are none of them primitive cases. They only occur in minds long schooled to doubt by the contradictions of experience. The primitive impulse is to affirm immediately the reality of all that is conceived.⁴³

Now, cases of detached sober rational assessment are cognitively sophisticated whereas the primitive mode of apprehension of something as real is not. James maintains that there is a general inclination to believe, and that we have to learn to distrust on the basis of contradictions in our experiences. It is only when contradictions and discrepancies are recognised that there is a need to resolve the felt cognitive tension.⁴⁴

A few decades later, Bertrand Russell examines James’ theory of unreflecting assent in The Analysis of Mind:

If this is correct, it follows (though James does not draw the inference) that there is no need of any specific feeling called ‘belief,’ and that the mere existence of images yields all that is required. The state of mind in which we merely consider a proposition, without believing or disbelieving it will then appear as a sophisticated product, the result of some rival force adding to the image-proposition a positive feeling which may be called suspense or non-belief – a feeling which may be compared to that of a man about to run a race waiting for the signal. Such a man, though not moving, is in a very different condition from that of a man quietly at rest. And so the man who is considering a proposition without believing it will be in a state of tension, restraining the natural tendency to act upon the proposition which he would display if nothing interfered. In this view belief primarily consists merely in the existence of the appropriate images without any counteracting forces.

There is a great deal to be said in favour of this view, and I have some hesitation in regarding it as inadequate. It fits admirably with the phenomena of dreams and hallucinatory images, and it is recommended by the way in which it accords with mental development. Doubt, suspense of judgement and disbelief all seem later and more complex than a wholly unreflecting assent. Belief as a positive phenomenon, if it exists, may be regarded, in this view, as a product of doubt, a decision after debate, an acceptance, not merely of this, but of this-rather-than-that.⁴⁵

Russell seems to agree with James that some simple forms of awareness entail a belief-like trust, and he mentions dreams and hallucinations as examples of such unreflecting assent. In cases where there is neither uncertainty nor doubt there is a naïve trust, whereas uncertainty and doubt are the products of rational considerations. Even if Russell is sympathetic to the idea that some states like dreaming involve an unreflecting trust, such states should not be conflated with fully developed beliefs.⁴⁶ It seems as if Russell comes very close to a plausible interpretation of Aristotle’s position. Rudimentary belief-like phenomena may occur in the absence of contradictory information, for example, in dreams, but full-fledged beliefs are the outcome of a rational choice of this over that, and such judgements require something more advanced than a form of unreflecting trust that is the outcome of a simple awareness of something.

7.1 Imagination Theories of Dreaming

There are a variety of views that reject the idea that dreaming involves a false apprehension of reality. As we have seen, appearing real and also trusting the dream to be real are prominent elements in many accounts of delusional dreaming (Aristotle, Spinoza, James, and so forth).

A number of theories that we may call imagination theories challenge the traditional idea of deception in dreaming. The label “imagination theory” of dreaming minimally entails that dreams are conceived as imaginations in the sense of being internal manifestations of a sensory character that are distinct from perception proper and are under voluntary control. It is often maintained that imaginations occur in the form of unfolding sequences of events or stories. Further, most imagination theorists stress that imaginations should not be conflated with other cognitive functions like belief. A central idea in many contemporary conceptions of imagination is that we can imagine arbitrary events, for example, that we travel to the planet Mars. However, it is sometimes supposed that acts of imagination also may seemingly emulate primitive forms of trust, for example, trust that we really encounter the things that we imagine. Exactly what this means will be discussed in detail below.

A quick comparison between the notion of “imagination” as used by modern imagination theories and Aristotle’s conceptions of phantasía (the faculty of imagination) and phántasma (sensory replication in any sense-modality) reveals a superficial resemblance, but Aristotle’s conception of imagination merely partially overlaps with contemporary conceptions of the term. According to Aristotle, dreams are perceptual remnants that emerge through phantasía. Sensory after-effects and re-activations (so-called phantásmata) occur under different cognitive circumstances, for example, in cases of misperception, in relation to thinking and memory, and in sleep in the form of dreams. Aristotle contrasts imagination (phantasía) with belief (dóxa), arguing that belief is involuntary whereas imagination, at least in waking, conforms to our will.⁴⁷ Elsewhere in De insomniis Aristotle says that dreaming is the work of the perceptual part, but belongs to this part in its imagining capacity.⁴⁸ So even if imagination is distinct from perception, they are closely related, which explains why imagination appears in a perception-like guise. One point of divergence between Aristotle’s and modern conceptions of imagination is that Aristotle’s notion of phántasma is object-oriented whereas modern theories of imagination often emphasise the narrative-like products of imagination.

However, there are at least two noteworthy correspondences between Aristotle’s conception and modern conceptions of imagination. First, both notions of imagination are regarded as strictly distinct from other cognitive attitudes such as belief. Still, imagination may under some circumstances superficially mimic something akin to a state of unreflecting trust or belief. Even so, it is quite difficult to determine to what extent modern imagination-theories are influenced by Aristotelian ideas of, for instance, non-doxastic belief-like phenomena or whether they are responses to Cartesian assumptions about intact belief-formation in the state of sleep.

However, to what extent modern imagination-theories oppose the tenets of the orthodox view varies – tenets such as (1a) that dreams appear in a way similar to the way perceptions appear during waking, or (1b) that dreams are percepts, and that (2a) dreams involve a misapprehension of reality or (2b) that dreams involve beliefs in the reality of the dream. I shall argue that some imagination theories fail to provide an account that accords with the widespread assumption that dreams in most cases are apprehended as real. Imagination theories that explain the dreamer’s absorption in a dream as the sort of immersion we undergo when, for instance, we read a story, seem better equipped to explain ambiguous states that are on the brink of delusional dreaming than they are to explain fully-developed delusional states.

As we saw, the orthodox view characterises dreams as perception-like experiences. Aristotle explains the perception-like nature of dreams by assuming that they are remnants of proper sense-impressions that no longer indicate the presence of real-world objects. In the modern discussions there is some disagreement on how to articulate the distinction between image and percept. However, the proponents of the so-called orthodox view quite often characterise dreams as perception-like experiences, but not necessarily perceptions, a description that refers to how dreams appear to the subject in the dream-state. The issue of how images relate to percepts remains controversial.⁴⁹ Part of the problem is the lack of standardised definitions of ‘imagination’ and ‘percept.’ In any event, a plausible theory should be able to account for the perception-like appearance of dreams, even if dreams are not considered to be ‘percepts’ according to very restricted applications of the term.

7.2 Fascination and Fictional Immersion

Sartre presents an intriguing theory that aims to show that the dreamer does not believe his dreams to be real events – the dreamer is rather compelled to attend to the dream by means of a narrow-minded consciousness. Dreaming is characterised by a spellbinding attention to unreal imaginations.⁵⁰ Sartre makes it clear that when we are caught up in imagination this is not a misapprehension of reality. According to him, the dreamer cannot apprehend the dream-image as real since he has lost the very conception of reality. This view is difficult to accept for a variety of reasons. For instance, if the notion of reality is lost, there seems to be no awareness of the experience as unreal either, which seems to make the experience neither real nor illusory. Yet Sartre clings to the idea that the dreamer is in some sense aware of the dream as an illusory image.

Sartre describes dream-consciousness as similar to the spell-bound state of our consciousness when we are absorbed in reading a fictional story: we buy into the story we are reading, but we are not mistaking it for real events. Colin McGinn presents a view that recalls of Sartre’s theory. McGinn calls this theory the “immersion theory of dreaming.” He writes:

Just as your mind cannot wander from your daydreams and expect them to proceed by themselves, so it cannot wander from your dream images – and the reason in both cases is the attention-dependence of the imagination. This explains the enthralling character of dreams, the single-mindedness of the dream state. It is not that dreams are somehow intrinsically fascinating so gripping that you cannot take your mind off them; on the contrary, they can be quite boring in the retelling. It is that they de facto have a monopoly on the attention. Since they are constituted by the attention, they are not the kind of thing from which the attention might wander. Their fascination for the dreaming consciousness is therefore an artifact of their constitutive nature, not a reflection of the narrative powers of their author.⁵¹

This idea reflects Sartre’s assumption that the reality perspective is absent in the dream-state. Thus, the dreamer is unable to be aware of the narrow-mindedness of the dream-state as opposed to the broader state of awareness that characterises waking consciousness. Yet Sartre maintains that the dreamer is not fooled by the illusory nature of the dream.

7.3 Do Dreamers Believe Their Dreams?

Sartre argues that no matter how caught up the dreamer is in his dreams, he has some rudimentary awareness that the dreams are mere images. In other words, he does not believe the dream. Sartre writes:

Here it is necessary to characterize the degree of belief of consciousness in the imaginary worlds, or if you prefer the “weight” of these worlds. Let us return to Mlle B …’s dream. The sole fact that the dream is given as a story should permit us to understand the kind of belief that we can attribute to it. But the dreamer instructs us still better, she tells us that she believed she was reading this story. What does she mean, if not that the story is presented to her with the same type of interest and credibility as that of a read story? Reading is a kind of fascination, and when I read a detective story I believe in what I read. But this does not signify in the least that I cease to hold the detective’s adventures to be imaginary. Simply, an entire world appears to me as imaged through the lines of the book (I have already shown that the words serve as an analogon) and this world encloses my consciousness, I cannot disengage, I am fascinated by it. This is the kind of fascination without positing existence that I call belief.⁵²

Sartre seems to argue that the enchanting fascination that characterises the dream-state should not be understood as an affirmative belief that what is dreamt is real. The dreamer rather accepts the dream as he might accept a story in a novel. We play along with the story in a sort of make-believe, yet we know that the dream-story is unreal.

Sartre may be right, given a particular interpretation: perhaps the dream exhibits no explicit or conscious claims of reality, that is, the dreamer does not explicitly think “This is real.” The dreamer may misapprehend the reality of the dream nevertheless. It is difficult to see how the state of fascination, in its most developed and compelling form, does not imply a misapprehension of reality. Sartre explains that what might appear as certain mental attitudes in dreams (for example beliefs) are really imagined fake imitations of real beliefs. But then, does not an ersatz-belief that the dream is real amount to the very same thing as real belief that the dream is real, namely, deception about the true nature of the dream?

Just like Sartre, McGinn maintains that the dreamer is not misapprehending his experiences as real events. He writes:

The attraction of this theory is obvious: it reconciles the image theory of dreams with the phenomenon of dream belief. While I am immersed in a novel, I am not under the strange delusion that the marks on paper are real events that I am observing; I know that I am only reading a book, not witnessing the events described in it. Nor do I mistake the images that form in my mind for belief-inviting percepts. Yet I am able to enter into the story to such a degree that my emotions may be stirred, rather as if I were witnessing these events. Perhaps even closer to the dream, when am I watching a film I do not confuse the images on the screen with real events; nor do I mistake my prompted imaginings for reality. Yet I may find myself so absorbed that my state of mind mimics real belief and feeling; I “enter into” the story. Similarly, in a dream I am not under the illusion that the images are percepts – I am implicitly aware (in some sense) that they are not – yet I am able to enter into the dream fiction in such a way as to become emotionally affected. I am not confused about the status of dream experiences; it is just that the dream images can draw me into a fictional world in such a way as to engage my cognitive and affective faculties. So engrossed am I by the dream story that I give my assent to it – or go into a state that is very similar to ordinary assent. Fictional immersion stimulates belief.⁵³

McGinn claims that we are not confused about the illusory nature of dream-experiences. Even if there is no genuine misapprehension of reality in dreaming, why does the dreamer not react appropriately in response to bizarre and incredible features of dreaming? McGinn explains:

I know very well that the actor on the stage is not about to stab the other actor, but I ‘believe’ that he is. I become absorbed in a novel in which a certain world leader has been assassinated, but I know very well that he has not. I am hypnotized into believing that I am a barking dog, but part of me knows that this is rubbish. The dreamer’s tolerance of inconsistency is therefore not some kind of preternatural irrationality or disregard for logic; it is simply the correlative of fictional immersion.⁵⁴

The tolerance for incredible elements is explained as an effect of playing along with the story; it is not a kind of acceptance that the dream is believable or real.

Nevertheless, McGinn seems to accept that dream-immersion may stimulate belief. But these are not really beliefs, they are more like quasi-beliefs.⁵⁵ McGinn writes:

We should not make the mistake of supposing that everything we call ‘belief’ fits some chosen paradigm of belief – say, assenting to a sentence when confronted by a sensory stimulus. Beliefs come in a great many forms everything from beliefs about perceived matters, to ethical beliefs, to theoretical beliefs, to religious beliefs (‘faith’), to dream beliefs.⁵⁶

And a few pages earlier:

I noted earlier that dream fear is not quite the same as real fear; it doesn’t have quite the clout of real fear. And dream belief is not quite as committed as ordinary belief; there is some kind of holding back or reservation about it. It is very hard to characterize exactly what this involves, but the point I want to make now is that the same kind of holding back applies to the emotions felt in ordinary fictional immersion. The belief and emotion of fictional immersion are quasi-belief and quasi-emotion (whatever this may ultimately come to).⁵⁷

McGinn seems to follow Sartre here when he claims that quasi-believing is something like going along with the story without strictly believing that it is true. The main weakness of this view is that it does not correspond to how people normally experience their dreams. For example, there seems to be no holding back when we are in the grip of horrifying nightmares. On the contrary, we psychologically respond to dreams as if the events dreamed of are really happening. Perhaps the reluctance to accept the idea of deception in dreaming relies on the assumption that the state of sleep disables belief-formation and that deception requires belief. However, to assume a kind of holding back seems to be a high price to pay given the almost universal experience that the dream-state in most cases deceives us (see note 1 above).

Jonathan Ichikawa, on the other hand, seems to argue that the dreamer is completely deceived regarding the reality of the dream, at least implicitly, but that this is not a belief, strictly speaking.⁵⁸ Ichikawa’s stance on deception depends on how the phrase “as if you are really there” is interpreted. He writes:

Lose yourself enough in your daydream, and you will feel, in some sense, as if you are really there. That’s not to say you falsely believe the contents of the daydream to be true. Our dreams in sleep are, on the imagination model, like that.⁵⁹

Ernest Sosa, along the same line of thought, introduces the notion of make-believe,⁶⁰ as a substitute for proper belief and just like Itchikawa he also seems to accept that deception does not imply belief.⁶¹

When something happens in my dream, reality tends not to follow suit. When in my dream I am chased by a lion, this poses no threat to my skin. No physical proposition about the layout of the world around me is true in actuality just because it is true in my dream. What about mental propositions about how it is in my own mind? Must any such proposition be true in actuality whenever it is true in my dream? No, even if in my dream I believe that a lion is after me, and even if in my dream I intend to keep running, in actuality I have no such belief or intention. What is in question is the inference from <in my dream I believe (or intend) such and such> to <In actuality I so believe (or intend)>.⁶²

In sum, the contemporary proponents of imagination-theories all circle around the same idea of being caught up in a dream and introduce a set of notions that are meant to substitute for proper instances of belief. The relevant notions include “fascination,” “immersion,” “quasi-belief,” and “make-belief.”

How then is imagination supposed to mimic belief? The idea seems to be that imaginations may simulate the occurrence of a variety of cognitive attitudes. So, when quasi – or pseudo – beliefs occur in dreams they are empty, ghostly replicas of the real thing. The difference between proper belief and imagined belief is as great as that between imagining the Eiffel Tower and perceiving the real Eiffel Tower. So when in the dream I take the horse I experience to be real, I do not really believe it to be real, I just imagine myself to believe it to be real.⁶³ Conversely, when I perceive the real Eiffel Tower I can form proper beliefs about the Eiffel Tower, for example, that it is real and not just some optical illusion. So, according to one interpretation of Ichikawa’s and Sosa’s accounts, “imagining x to be real” as this occurs in dreaming has the same practical consequence of deception as-if this was really believed, even if it is not. Let us now take a closer look at some theorists’ reluctance to accept that the dreamer is deceived with regard to the reality of the dream.

As we have seen, the assumption that dream subjects are never completely deceived about the reality of the dream conflicts with ordinary experiences of dreaming. Even so, the imagination theory seems well suited to explain borderline cases between delusional awareness and lucidity, that is, ambiguous awareness of a dream’s reality, such as the awareness that there is something strange about the dream that is difficult to articulate while we remain in the dream-state. For example, I have often dreamed of interacting with people who have passed away. In such dreams, at some points, I have felt that something is wrong, as if a part of me knows that the current dream contradicts known facts. On a charitable reading, this ‘feeling that something is wrong’ might correspond to McGinn’s description that something is holding us back in dreaming. As mentioned, Sartre and McGinn deny complete deception and assume some degree of lucidity in every case of dreaming. However, the occurrence of borderline cases between delusional dreaming and lucidity does not rule out dream-states in which the subject is completely deceived about the illusory nature of the dream.⁶⁴ So, are people mistaken when they say that they are deceived by the realistic appearance of dreams? If Sartre and McGinn are right, people merely say that they mistake dreams for real events, but they do not really.

Let us sum up some of the points made by proponents of imagination theories of dreaming and their stance on the question of delusional dreaming. (1) Are we deceived (or deluded) about the reality of the dream? Sartre: No, not really, but we are immersed in the dream by the narrow-minded consciousness that constitutes dreaming. Something in us is aware of the dream as an unreal imagination. McGinn: No, there is a holding back of the kind we experience when we are immersed in fictions. Yet, immersion can stimulate belief-like attitudes. Ichikawa: Yes, we can be deceived, it is as if we really are at the place we dream about, but we do not strictly speaking believe it. Sosa: Yes, we respond to the dream as real without believing it, because in dreaming we are unable to believe that the dream is real by means of proper belief.

One detail that divides modern imagination theories is the issue of whether deception occurs at all. Now, all the discussed varieties of imagination theories presuppose that beliefs are ruled out regardless of whether the dreamer is deceived. Sartre and McGinn reject the possibility of deception whereas Ichikawa and Sosa seem to accept a form of deception that is not based on belief. The latter position, apparently involves a kind of immersion in dreams that approximates the non-doxastic view that has been attributed to Aristotle in this paper.

7.4 Owen Flanagan: Delusional Dreaming through Absent Metacognition

Owen Flanagan, a philosopher in the contemporary analytical tradition, presents an account of delusional dreaming that comes close to Aristotle’s view that the kind of unreflecting trust that is characteristic for dreaming is a consequence of the temporary shut-down of a set of higher cognitive functions. He writes:

Another way in which dreams differ from wakefulness is this. Many philosophers, thinking they follow Descartes – take for example, Russell’s claim that the awake and dreamed thoughts of the ruined church are ‘intrinsically indistinguishable’ from each other – think that the fact that often we cannot tell we are dreaming when we are dreaming is because dreams seem as real as real can seem. One reason dreams seem so real is related to the point just made: our metacognitive powers are typically turned way down in dreams and thus so are our judgemental capacities to perceive incongruities, uncertainties, and discontinuities as odd while they are occurring. Perceiving that such things are odd is hard, since perceiving them as odd requires layering thoughts, for example, having the thought that one is flying while also having the thought that people can’t fly. We understand from a neurophysiological perspective why we are better able to think metacognitively when we are awake than asleep. But in any case, the fact that dreams seem so real while we are dreaming that we cannot tell that we are not awake does not imply that dream mentation is just like awake mentation in terms, say, of vivacity or along other phenomenological dimensions.⁶⁵

Flanagan highlights an important detail: we can be deceived about the reality of dreams even if they are very unlike ordinary perceptions we have in the state of waking. For example, it may be the case that dreams are less vivid than perceptual experiences and that they typically deviate from perceptions in a number of distinctive ways. Moreover, dreams often display unrealistic content that does not correspond to waking experiences (cf. section two on different senses of ‘appearing real’ above): it is as if we are unable to detect vital differences between dreams and waking consciousness while we remain in the deficient mode of cognition that is characteristic of the dream-state. Differently put, we are not necessarily deceived by the dream because it mimics waking experiences to some degree; rather, the dreamer mistakes the dream for reality because of the altered cognitive state he is in, not because the dream in itself is indistinguishable from waking experience.

Flanagan observes that in order to be aware of (1) incongruities (e.g., having a conversation with a duck on the campus lawn), (2) uncertainties (a person may look like a known person but is at the same time felt to have some different or indeterminate identity), and (3) discontinuities (my dream-self undergoes rapid shifts in location – in one moment I am in Europe another moment I am in the Americas), as odd, we have to have layered thought. In other words, we have to have thoughts about thoughts (cf. Brito’s view of the role of the intellect in delusional dreaming). In the dream-state we seem to be unable to assess content in a rational way through the firmly held beliefs we have established in the state of wakefulness. Thus, it is the absence of metacognitive assessments that explains why the subject mistakes the dream for reality.

Flanagan’s explanation of delusional dreaming comes close to Aristotle’s account because the delusional appearance of reality is a mere effect caused by the inactivity of certain supervising cognitive capacities. However, unlike Aristotle, Flanagan explicitly discusses the inability to detect irrational and atypical events and unfolding of events. Another important difference between the two accounts is that Aristotle stresses the similarity between dreams and perception, whereas Flanagan emphases their differences.

7.5 Jennifer Windt: Unreflecting Trust, Deception, and Adoxastic Dream-States

Jennifer Windt offers a very rich account and I shall only consider her take on dream-deception in relation to the notion of unreflecting trust. Just like Flanagan, Windt makes a distinction between deception based on phenomenal resemblance and deception due to cognitive impairment. She introduces the notion of doxastic situatedness, and beliefs are said to be paradigmatic cases of doxastic situatedness. She then explains:

Doxastic situatedness refers to exactly those attitude types that carry with them particular epistemic commitments. If I believe or affirm that p, I am convinced or firmly assume that p, and so on, then I am thereby committed to the truth of p. This commitment influences how I subsequently reason or act on p. If I am strongly convinced of a particular theoretical claim, I will defend it against objections; and if I clearly remember having left my keys in my purse, I will search for them there when needed, which deplorably is not always the most direct path to success.⁶⁶

Windt’s main point seems to be that some dreams exhibit no doxastic situatedness at all, not even in the form of a rudimentary unreflecting trust. Such dream-states are considered to be doxastically undetermined and are adoxastic in nature. In fact, Windt argues that a majority of dreams are of this adoxastic character. She writes:

A new way of saying that dreams are deceptive, then, is to say that dreaming is a state of doxastic disorientation: dreams mislocalize us in our doxastic framework by preventing access to long-standing beliefs or memories, but also by temporarily enticing us to endorse new ones. Dreams are, literally, misleading experiences. In the Cartesian scenario of dream-deception, the dreamer believes or affirms propositions such as ‘I am awake,’ ‘I am really holding this piece of paper in my hands and not just dreaming that I am,’ and so on. The dream self is now doxasticially situated toward the content of these propositions in the manner of believing of affirming their truth. Again, not all dream cogitations involve doxastic situatedness, and only a subgroup that do will affect how one is situated toward the fact that one is now dreaming. If dream beliefs can occur in isolation, as argued in section 9.6, then believing, in a dream, that my mother has grown a beard does not entail that I thereby take myself to be awake. Other dream-beliefs will undermine one’s prior doxastic position without substituting a new one. One can be doxastically situated toward the proposition I am awake’ by simply doubting its truth. In this case one will no longer believe that one is awake, but one will also not believe that one is dreaming. What is essential for dream deception, then, is being doxastically situated toward a certain type of content, expressed in propositions about one’s current state of consciousness and the reality of ongoing experience, in a manner that is at odds with the actual truth or falsity of the respective propositions.⁶⁷

Windt suggests that dreams are deceptive in a special sense because they do not stimulate adequate doxastic attitudes in the way that waking experiences do. Nevertheless, she also suggests that a large class of dreams lack any doxastic element – they do not include even rudimentary forms of unreflecting trust – and therefore these cases do not involve deception, strictly speaking, because deception requires at least some minimal degree of doxastic situatedness. So, a mistake about the reality of a dream requires that this reality is affirmed, no matter how simple such affirmations are.

Windt challenges a central element in the Aristotelian account of the mistaken reality of dreams, namely, that unreflecting trust automatically results from the simple awareness of a sense-impression. Instead, she argues that an awareness of p as such does not imply blind trust in p because of the absence of opposing considerations. She writes:

What about intermediate states between doxastic indeterminacy and doxastic situatedness? A possible objection to my account is that dreams involve unquestioning assent rather than full-fledged belief and that this is enough for deception. For instance, Reed while granting that ‘ordinary dreams involve no considered opinion or strongly appraisive belief’ (Reed, 1979, p. 45), argues that the concept of taking for granted should be applied to dreaming. Here “the dimension of appraisal is minimal … and the possibility of deception is assured” (pp. 43–44) since one can take something for granted that is false. According to Reed, this is enough to vindicate the traditional view: “Ordinary dreaming involve belief, in the perhaps extended or uncommon sense of ‘taking for granted’ ” (p. 44).

To this, I would respond that taking for granted, or unreflected assent, may well be a precursor to more sophisticated forms of belief. The distinction between doxastic indeterminacy and doxastic situatedness is intended to be not sharp but gradual. Taking something for granted in a dream – for instance, that this is not a dream, but is really happening – is a simple and unarticulated form of doxastically situating oneself toward the fact that one is now dreaming. In doing so, one fulfils the condition for the simplest form of deception. What I would question, however, is that we should ascribe even this simple form of belief to dreams in the absence of any positive evidence. Rather, it might be useful to distinguish such cases of taking for granted from doxastically indeterminate dreams. The concept of doxastic indeterminacy makes room for the possibility that experience does not automatically invite beliefs and does not per se involve even a primitive form of belief such as taking for granted.⁶⁸

Windt’s reference to adoxastic states may superficially appear to involve a quite non-committing position. However, it is not clear what the claim that most dreams are adoxastic amounts to. For instance, we may ask (1) are there adoxastic dream-states, and (2) if there are, how many dreams have this adoxastic character, and (3) do such states involve any kind of deception? I shall argue that the average awareness of dreams involves a form of unreflecting trust in the reality of the dream. The proposed view challenges Windt’s position which claims that most instances of dreaming lack any doxastic element no matter how rudimentary it is assumed to be. In my view, if the dream seems to take place in the real world and this appearance is not questioned in any way, then there is deception about the reality of the dream. On the other hand, rudimentary forms of unreflecting trust do not imply articulated beliefs like “This is real” or “I am not sleeping.” Dreams may be taken to be real, not necessarily through a conscious judgement that what is experienced is real, but indirectly by viewing events as if appearing in the world and not noticing that one is asleep. This implicit apprehension of the dream as real is sufficient for deception regarding the reality of the dream.

Let us examine Windt’s assumption of alleged adoxastic states in dreaming. What does it mean to be aware of a dream yet remain neutral to how the dream appears? Is it a disengaged attitude of “going through the motions”? Is it a completely neutral apprehension of a dream with regard to its realistic qualities? Is it like watching a movie but not responding to what is seen with actions or emotions? Or does it simply mean that the realistic features of a given dream are not reported upon awakening? For example, would my dream of my friend D reading on the campus lawn qualify as such an adoxastic state? I do not apprehend the dream as real in any conscious significant way at all. While dreaming, I am just completely caught up in the situation here and now. I did not explicitly think: “This is real,” and I did not disbelieve that it was an occurrence in the world. It seems right that there is no deception of reality in delusional dreaming that matches sophisticated cases where we sincerely wonder: “Are my sensory appearances deceiving me?” In my dream there was not even any consciously articulated conception of reality, only a tacit, intuitive sense of reality. In fact, deceptions of reality do not seem to require any articulated concept of reality. Imagine a cat being deceived about chasing an unreal mouse while dreaming. If a cat can be deceived by a dream it seems unreasonable to suppose that deception requires the cognitive resources to articulate conceptually sophisticated attitudes about reality as “reality.”

The dreamer is not only deceived about the true nature of the things dreamt of, the dreamer is also deceived about the cognitive states that monitors the events in the dream. This move turns things back to the endorsed interpretation of Aristotle and suggests in just how deflated a sense ‘unreflecting trust’ ought to be understood. It is a tacit unreflecting trust that things are really happening in the world, not in the form of articulated doxastic attitudes like “This has to be real,” “This really appears real,” or “I am truly convinced that this is real,” but rather in an unquestioned intuitive yet prominent sense of reality that is not the result of any top-down processing.

It seems like Windt’s discussion disregards the detail that dreams can appear to be real and that appearances play a vital role in accounts of why dreams are apprehended as real. Now, to uncritically take dreams at face value is not a matter of intellectual processing. But why then do dreams present themselves as real rather than neutral or unreal? Here is one answer. Perception paradigmatically gives us immediate and intuitive access to the world. Sensory imaginings of the kind we entertain in the state of waking do not really indicate the presence of real-world objects, yet their existence is parasitic on proper perceptual states. Dreams along with imaginations mimic the characteristic format of perception that things take place in a world and thereby emulate the realistic appearance of perception. The illusion of reality that is inherent in imagination is less prominent when we perceive the world around us and conversely the illusion of reality in imagination and dreaming is increased the more perception, and other higher cognitive functions are deprived.