More than a Critic of Locke and Berkeley: Baxter on Dreams and Imagination

I. Baxter’s Account of Dreams: Phenomenology and Knowledge

Baxter’s theory of dreams is presented in his An Enquiry into the Nature of the Human Soul (1733; 3rd ed. 1745; hereafter Enquiry), particularly in a long essay on the origin of dreams. This essay, originally published as Section VII of the Enquiry, bears the rather cumbersome title An Essay on the Phaenomenon of Dreaming, wherein is shewn from the INERTIA of matter, and the nature of mechanism above explained that this appearance cannot be the effect of mechanism or any cause working mechanically; and thence that it must be the effect of a living, designing cause. The several hypotheses for solving this appearance mechanically, particularly examined, &c. (hereafter Essay on Dreams or ED; all references are to volume two of the third edition of the Enquiry). While the title only hints at his conclusion, Baxter arrives at what he takes to be the best—indeed, only coherent—hypothesis concerning the origin of dreams through an elaborate critique of alternative explanations: that dreams are produced by our mental acts, by the physical processes of our bodies, or by mere chance. To the consternation of both his contemporaries and later readers, his theory posits as the source of all our dreams separate, finite minds of questionable moral status but great power—that is, ghosts.²

In this paper, I aim to contribute to a fuller appreciation of the significance of Baxter’s theory by first setting out the richness and perceptiveness of his account of dreams, including its phenomenology, before going on to underline its main philosophical appeal—namely, that it draws a strong distinction between imagination and dreaming, and as such, provides a serious challenge to his predecessors. This should shed more light not only on the originality and the unprecedented depth of his conception of dreaming, but also on the extent to which it is continuous with other, more standard theories of dreams of his time. The phenomenological feature that Baxter emphasises most throughout his discussion is the passivity of the dream experience with respect to the scenes presented and the events happening to our dream selves. While the passivity of dreaming has often been implied by those who hold that dreams can be deceptively realistic, Baxter makes this point more explicit and introduces an important distinction.

By ‘passivity’ Baxter means not only the absence of the feeling of producing the dream experiences we are having (the sense of mere involuntariness, or perhaps better, non-voluntariness), but also the more positive feeling of passivity or involuntariness, in the sense of experiencing something being imposed on us, without or even against our will (see, e.g., ED: 13–14). Interestingly, Baxter thinks that we are also active in our dreams, in that we form conscious volitions—real volitions (to produce dream actions), which our dream selves form only because the scenes, as well as the actions of apparently other beings, exhibited to us feel real. As he claims, ‘there is the same difference and distinction of consciousness, betwixt what the soul itself doth, and what the persons in the vision seem to do’ as in real life (ED: 35, note p; see also ED: 22).

Unsurprisingly, the passivity of dream experiences contributes to their feeling as if they were real. Dreamers, as he argues, are ‘just as passive in receiving these impressions as it would be in receiving the like impressions from external objects […] when broad awake’ (ED: 13–14). While involved in the dreaming, our minds are ‘thoroughly possessed […] with a belief of the reality of the things represented’ (ED: 15, note e). A crucial aspect of this realistic phenomenology is not merely that we lack conscious control over our dream contents—which perhaps would be compatible with the view, rejected by Baxter, that our minds involuntarily cause them—but that many dream experiences are felt as if imposed on us by external factors or agents. The terminological choice of ‘possessed’ is, of course, not entirely metaphorical here.

Another part of this lifelike appearance is that we do not merely dream about our own actions and those of other agents or spirits being imposed on us, but also about all the ‘circumstances of time, place, company’ (ED: 34, note p), being fully detailed, as if occurring in a fully determinate reality. Moreover, a dream is not a ‘dead’ or ‘blind’ unfolding of a predetermined chain of events, like a script, but a living, spontaneous, and interactive experience. According to Baxter, this feature of dreams indicates that other spirits are causally involved in the creation of our dream world. Otherwise, the dream experience would feel much more static, artificial, and under our conscious control than it actually does.

A further salient feature of dreams (now viewed from a waking perspective) is their recurring strangeness and chaos, which Baxter emphasises alongside what we might call ‘signs of rationality or design’—aspects of dreams, he strongly believes, the standard materialist explanations of his time cannot account for (for details, see Bartha 2025). Not only does Baxter point to various inconsistencies both within the dream experience itself and in relation to our waking experience, but he adds that the unexpected and puzzling nature of dreams often reaches a level of creativity surpassing that of even the most imaginative artists (ED: 9, 35–36).³ Baxter also notes that dreams exhibit a phenomenon we might call ‘unfounded familiarity’, where new things, strangers, or surprising situations appear familiar to us, and we are completely untroubled by going along with them as normal—much to the surprise of our waking selves recalling these totally unexpected dream events (ED: 206–10, 229–30).⁴

For Baxter, this observation reaches beyond mere psychology, hinting at the possibility of ‘instantaneous knowledge’. In such dream situations, we often (feel we) know a great deal about these ‘new’ people and circumstances without ever having learned any facts about them (ED: 210). Such cases, he argues, do more than raise the logical possibility that our minds may operate in ways unavailable to us in waking life; they in fact bear witness to the remarkable cognitive powers that the soul possesses independently of the body and the physical processes underlying sensation (i.e., in the afterlife). Dreaming, therefore, reveals a cognitive operation considered higher—‘more perfect and ready’—than our ordinary waking way of gradually acquiring knowledge through the labourious processes of observation, comparison, inference, reasoning, and memory (ED: 220–21, and esp. 231).⁵

In discussing the instantaneous nature of dream knowledge, as in many other places in the essay, Baxter draws on other thinkers—especially ancient authors—to argue as a ‘moral certainty’, accepted by many philosophers as well as by common sense, that we can derive knowledge from our dreams: about future events, speculative truths, or ourselves.⁶ He cites Lucretius, for instance, who held that dreams can reveal our ‘own secret villainies’ (ED: 225; see also 226–28). Moreover, not only does Baxter revive certain ancient theories of dreams—if not anticipate Freudian psychoanalysis (which, of course, is fundamentally at odds with his commitment to the transparency of the mind)— he also inverts the dominant epistemological assumptions of his time. Whereas early modern philosophy typically devalues states such as dreaming and imagination—regarding them as secondary, derivative, parasitic, or philosophically unrevealing in comparison to waking perception and thought—Baxter treats dreaming as epistemically significant. Despite the (literally) spooky causal story of external spirits interfering with our minds and brains, dreaming, for Baxter, offers insight into how minds (human souls and other spirits) are capable of interacting and functioning on their own terms—often more freely than in waking life, where cognitive activity is constrained or distracted by sensory input. In addition to his claim that dreaming displays unseen levels of creativity, I suspect it is precisely this revaluation of dreaming’s epistemic significance that accounts for Coleridge’s deep resonance with Baxter’s account.

II. Dreaming and Imagination in Baxter and His Predecessors

Baxter’s characterisation of the salient features of dreaming is not only interesting in itself, as it reveals how carefully he reflected on its phenomenology and its broader significance for our cognitive capacities. Moreover, when combined with his conclusion about the spiritual origin of dreams—that they are caused by external, immaterial agents rather than by our own minds or bodily processes—this distinctive phenomenological analysis enables him to categorically distinguish dreaming from imagination and other related phenomena, such as recollection. Early on, Baxter promises to explore the ‘affinity between possessing the fancy in sleep, or dreaming, and possessing it while awake’ (ED: 5). More precisely, however, his aim is to bring out the fundamental disparity between them. Indeed, he makes clear that the closer analogy lies between dreaming and waking possession, rather than between dreaming and ordinary imagination. In terms of the desideratum of a more clearly delineated faculty psychology, Baxter’s account can thus be seen as a marked improvement over those of his predecessors.⁷

According to Baxter’s immediate predecessors, the faculty of imagination is exercised in two forms—what we might call ‘waking’ and ‘sleeping fancy’—with ‘waking fancy’ typically encompassing both remembering and imagining. Thus, thinkers such as Hobbes, Locke, Berkeley, and Addison generally treat imagination (as well as memory) and dreaming as different manifestations of the same faculty: the representation of absent objects in a quasi-sensory manner.⁸ On this view, both dreaming and imagination are derivative of sensory perception. These operations fall under the same faculty insofar as they depend on sensory input and our memory of it, which is then reproduced or reconfigured by the mind or by the relevant bodily and brain processes, whether one is awake or asleep.

Accordingly, there is no categorical distinction among these operations but only in their relation to perception, which, being tied to an actually present, external object, is both conceptually and temporally prior to imagination, memory, and dreaming. Between imagination and memory, by contrast, the differences are a matter of degree—for instance, acts of remembering may differ from imagining in vividness, detail, or accuracy of reproduction. Dreams, for their part, may be more vivid than waking imagination due to the concurrent ‘silence of the sense’, as Hobbes puts it in Leviathan (I.2.5). Yet Hobbes insists that we imagine ‘as well sleeping as awake’, retaining ‘an image of the thing seen, though more obscure than when we see it’ (I.2.2). For Hobbes, the ‘imaginations of them that sleep’ (I.2.5) are simply ‘decaying sense’ (I.2.2), and are thus equally dependent on—if not reducible to—memory (I.2.3). The only structural difference lies in the direction of the motions involved: ‘our dreams are the reverse of our waking imaginations; the motion when we are awake beginning at one end, and when we dream, at another’ (I.2.6).⁹

As I will show in the next section, Berkeley shares the main contours of this picture, though in a fully mentalistic key. Locke, by contrast, not only subsumes dreaming under imagination but goes even further in downplaying the phenomenological distinction between these cognitive operations. Famously, he writes: ‘if our dreamer pleases to try, whether the glowing heat of a glass furnace, be barely a wandering imagination in a drowsy man’s fancy, by putting his hand into it, he may perhaps be wakened into a certainty greater than he could wish, that it is something more than bare imagination’ (Essay IV.xi.8). For Locke, dreaming is not merely a form of imagination with a similar causal origin, as Hobbes suggests—both being concerned with quasi-sensory ideas of absent objects produced by the mind from prior perceptions. More strikingly, contra Descartes, Hobbes, and later Berkeley, Locke holds that dreaming is phenomenologically closer to waking imagination than to sense perception or realistic hallucination. In dreams, we lack not only the vividness and causal efficacy of genuine sensory experience—such as pain—but also, ‘for the most part’, exhibit the ‘extravagant and incoherent’ or ‘frivolous and irrational’ qualities that, he claims, contrast sharply with the ‘perfection and order’ of our waking thoughts (Essay II.i.16; see also II.xix.4). According to Baxter, this is a gross mischaracterisation of the phenomenology of dreams.¹⁰

For Baxter, as for Descartes and the Cartesians, imagination (including recollection) is a faculty of dual ontological status.¹¹ As he clarifies in his essay against Berkeley—which follows the Essay on Dreams in volume two of the Enquiry and bears the typically informative title Dean Berkeley’s Scheme against the Existence of Matter and a Material World Examined, and Shewn Inconclusive (hereafter ‘Berkeley’s Scheme’):

Imagination, as it is the perception of a picture, shews not only that the soul is immaterial, but that it is united to a material sensory [organ], where the picture is impressed, and to which it applies for the perception of it; or that matter exists. (Enquiry II.282)

Like Descartes, Baxter holds that the mechanism of imagination already points toward the reality of an external world, and thus has the potential to ‘overturn D. B.’s [Dean Berkeley’s] scheme’ (Berkeley’s Scheme, Enquiry II.282). In a sense, the bodily basis of imagination, ‘the sensory [organ] where the impressions are made’, can itself be called imagination. A similar ontological structure applies to dreaming. In the same essay (Enquiry II.236, note a), he makes it clear that all dream theories, including his own, presuppose the existence of matter: ‘all these ways still suppose the real existence of matter, in supposing both a sensory [part in the brain] and objects [i.e., the brain images] acting upon it’.

But when we speak more precisely about the mental faculties of imagination and dreaming—and how the relevant parts of the body, especially the sensory region of the brain, are exercised to produce their respective visions—a stark contrast emerges between the two. According to Baxter, this contrast concerns not only their ultimate causal source, but also the nature of their operation. The conscious and voluntary activity essential to imagination is conspicuously absent in the case of the scenes and happenings of our dream world. Imagination, for Baxter, consists in the mind’s ‘[…] own active power of voluntarily joining ideas together, without objects ab extra to cause them’ (ED: 34, note p). Crucial to his broader argument for the spiritual origin of dreams, as well as to his philosophy of mind more generally, is the view that the mind ‘coupl[ing] together ideas by this active power, without being conscious of its own workmanship, is as little to be conceived’ (ED: 34, note p). Defined as an essentially conscious activity, imagination stands in clear, categorical contrast with dreaming, which is not only experienced as passive (except for our own volitions) and thus as if real, but is, in keeping with the transparency and self-evident authority of self-consciousness, in fact caused and imposed upon us by other spirits.

Accordingly, Baxter frequently emphasises the passivity of dreams in contrast to the activity of imagination. The same contrast, then, holds between memory and dreaming. However, since the two are closely intertwined, it is worth clarifying the precise role memory plays in dreaming. Like many of his predecessors, Baxter notes that one precondition for dreaming is a temporary forgetting—especially of recent waking experiences. Without this forgetfulness, our dream experiences would appear obviously out of sync with immediate memory and thus fail to convince us of their apparent reality (ED: 8, and elsewhere). At the same time, Baxter—modifying the Hobbesian picture—acknowledges that memory remains essential for supplying the raw materials of dreams: the residual traces in the brain, which are then rearranged in often novel and creative ways. Yet crucially, unlike Hobbes, Baxter insists that these dream images are not reassembled by our own mental faculties or bodily processes, but by external spirits acting upon us.

Nevertheless, Baxter maintains that remembering is categorically distinct from dreaming, since it is an essentially active mental operation—except in cases of involuntary or triggered recollection (such as those evoked by a madeleine, in Proust’s famous example), or what he calls ‘musing’, in which the mind merely reacts to an initial sensory or affective prompt.¹² In ED: 215–18, Baxter—following Locke more explicitly than Hobbes—defines memory as a renewed impression or perception, accompanied by a secondary or concomitant awareness that the original impression has been experienced before. While memory involves a passive bodily component, insofar as the original perception is retained in the brain, the act of remembering itself is active: the mind initiates the retrieval of a particular memory trace and sustains attention upon it. This makes memory, unlike dreaming, a self-directed activity of the mind.¹³

As I have tried to show in this section, Baxter believed that dreaming—despite its ontological similarity to imagination and its reliance on memory, and hence on prior perceptions for some of its basic materials—is best understood as a sui generis cognitive process. By this I mean simply that its essential and defining operation cannot be explained in terms of sensory perceptions or images being retained, renewed, and/or recombined through some mechanical process in the body or through the mind’s own initiative. As he summarises his position in the essay against Berkeley, ‘new impressions [by external spirits] are made upon’ the sensory part of the soul, and thus upon the mind (Enquiry II.236, note a). Dreaming is therefore reducible neither to the ideas of imagination produced by our own minds, nor to waking perceptions, which are caused by external objects through God’s causal mediation. With his distinctive account of both the phenomenology and causal origin of dreaming, Baxter transcends this false dichotomy and, by categorically distinguishing dreaming from the faculties of perception and imagination (including recollection), offers an arguably more fine-grained faculty psychology than his predecessors such as Hobbes, Locke, and Berkeley.

III. Challenging Locke and Berkeley’s Conception of Dreaming

There is, however, a deeper theoretical payoff to Baxter’s account of dreaming. His sharp contrast between imagination and memory on the one hand, and dreaming on the other—as causally, phenomenologically, and epistemologically distinct cognitive capacities—brings to light certain philosophical tensions in his predecessors, tensions that had been obscured by their failure to differentiate these functions in clear-cut terms. Baxter capitalises on these tensions to argue that his theory of their spiritual origin provides the only coherent explanation of dreams. He also appears to have these difficulties in mind when seeking to discredit Berkeley’s idealism as a form of radical scepticism, incapable of distinguishing dreaming from reality.

Indeed, thanks to his finer-grained faculty psychology, Baxter is uniquely positioned to highlight a point that, surprisingly, is rarely addressed by commentators: both Locke and Berkeley largely ignore the question of what causes our dreams. As a result, they also fail to account for how their (likely) view—that dreams are imaginings, or quasi-sensory states generated by our own minds—can be reconciled with their broader commitments concerning consciousness and mental activity. In particular, they offer no explanation for the striking phenomenological fact that dreams appear passive or involuntary—seemingly produced not by the dreamer’s own volition but by something external to it.¹⁴

The challenge, then, is twofold. First, despite the phenomenology suggesting otherwise, both Locke and Berkeley appear to classify dreaming as a mental activity—as a function of our own minds and, indeed, of our wills—on a par with ordinary acts of waking imagination or deliberate recollection. This stance conflicts with their (correct, in Baxter’s view) commitment to the infallible authority of self-consciousness, which, in the case of dreaming, clearly testifies to the contrary, indicating that dreams as passive rather than self-generated states. If, as I have suggested, Locke seeks to deny the essentially passive character of dreams, then, in Baxter’s eyes, he could scarcely be more mistaken.

But even if we set aside this disagreement about the phenomenological datum, Locke’s causal story remains difficult to reconcile with his broader commitments regarding the mind—particularly his views on mental transparency and the grounding of personal identity in consciousness and memory. If the mental activity that produces the dream scenes we perceive is construed as an operation entirely inaccessible to introspection, then Locke is effectively forced to abandon his strong commitment to the mind’s transparency to itself. Ironically, he seems to rely on that very principle in arguing against the Cartesian claim that the soul thinks essentially and thus continuously, even in states of deep or dreamless sleep. By contrast, Locke insists that what does and does not occur in sleep—dreaming included—falls under the authority of transparent self-consciousness (see especially Essay II.i.10–14). Even more famously, he appeals to this same transparency in his argument against innatism, claiming that ‘to imprint anything on the mind without the mind’s perceiving it seems to me hardly intelligible’ (Essay I.ii.5).¹⁵

Moreover, Locke appears to be forced to concede that the mental activity responsible for producing dream content must belong to someone other than the waking person who later remembers the dream but has no recollection of having generated it. If our consciousness or memory cannot be traced back to the mental operations that gave rise to the dream scenes, then, on Locke’s own terms, it must be another person’s volition that is responsible for them. This situation closely parallels Locke’s well-known thought experiments involving alternately waking and dreaming subjects, which amount to two distinct consciousnesses inhabiting the same body or soul, yet entirely unaware of each other’s experiences, and thus ‘as distinct as Socrates and Plato’ (Essay II.i.11–12, 15; see also II.xxvii.16, 21, 25). Not only is this scenario counterintuitive, but Locke himself presents it as a difficulty in the context of his critique of Cartesian assumptions about the unity of consciousness. Baxter discerns an even more troubling implication: not merely two temporally distinct persons occupying the same body or soul at different times, but two distinct persons—one actively producing and one, unaware of the other’s activity, passively experiencing the dreams—coexisting within the same mind at once. Baxter is understandably puzzled: ‘how could the soul, upon Mr. Locke’s own principles, form to itself in sleep a scene of our waking actions and thoughts, and the man be still ignorant of it, without being two distinct persons?’ (ED: 183; see also ED: 18–19, 51–54, 159–61).¹⁶ These tensions have gone largely unacknowledged—perhaps even deliberately swept under the carpet—due to the uncritical assimilation of dreaming to imagination, a Hobbesian inheritance that Locke, as well as Berkeley, seems to have taken for granted.

Even though Baxter more explicitly targets Locke with such arguments in his Essay on Dreams, and only hints at similar problems in his essay against Berkeley, Berkeley’s case is equally, if not more, striking.¹⁷ First, Berkeley faces—but never directly addresses—the obvious difficulty of how to accommodate dreaming within the dichotomy inherited from Hobbes and Locke between passively (or involuntarily) received sense perceptions and active (that is, voluntarily produced) imagination. While Berkeley often lists dreams alongside ideas of imagination (as in the Notebooks 823 and 843, with 843 also adding reveries, or daydreamings), there appears to be no other place for them within the threefold classification of ideas—of sense, reflection, and imagination—that he officially endorses in Principles of Human Knowledge (PHK 1). The implied view, therefore, is that dreams (and other involuntary forms of mental imagery) are essentially like imagination and are consequently caused by the perceiver’s own mind. Similarly, in Three Dialogues (DHP: 235), where both are contrasted with ideas of sense caused by God, the implication is again that ideas of imagination and dreaming share the same origin in ourselves—so much so that the implication is that dreams may simply be regarded as a form of imagination, what Baxter terms ‘sleeping fancy’. The obvious difference—that dreams are ideas of imagination produced unconsciously or involuntarily—is never explicitly acknowledged, let alone addressed, by Berkeley.¹⁸

This oversight is especially regrettable, since it obscures several apparent tensions within Berkeley’s own system. One such issue concerns his commitment to the mental transparency thesis. As Stuart (1997: 125) observes, there is no indication that Berkeley ever entertained the possibility of unconscious mental operations; on the contrary, he seems to presuppose a version of mental transparency. In the New Theory of Vision, for example, when rejecting the claim that visual perception involves imperceptible lines and angles, Berkeley explicitly insists that ‘everyone is himself the best judge of what he perceives, and what not’ (NTV 12). In this respect, Berkeley encounters the same difficulty as Locke; yet the problem is more severe for Berkeley, since—unlike Locke—he embraces the phenomenological passivity of dreaming.¹⁹ He appeals to dreams (and related phenomena) precisely to show that no recourse to mind-independent objects is required to explain experience. The subjective indistinguishability of dreams from waking perceptions demonstrates that we can be ‘affected with all the ideas we have now’, even with ideas ‘produced always in the same order we see them in the present’ (PHK 18). If dreams were not experienced as ostensibly passive, there would be an immediate and obvious way to distinguish them from waking ideas, undermining his point.²⁰

Of course, Locke and Berkeley could be defended here, and there are several ways in which they might respond to the challenge posed by Baxter’s strict demarcation between dreaming and imagination. In particular, unconscious mental operations or causal processes might be accommodated within the Lockean or Berkeleyan framework, provided that the ideas or perceptual objects generated by these processes are themselves conscious. What ultimately matters for both thinkers is that the outputs—not necessarily the underlying operations—are manifest to the mind that experiences them. On this view, the mind’s transparency may pose less of a difficulty than Baxter makes it out to be, since this model would cohere with Locke’s and Berkeley’s presumed accounts of the generation of dream ideas by a mind that is transparent with respect to its ideas, but not necessarily with respect to all its operations.

Their commitment to mental transparency might even be further qualified. Hatfield (2011: 367, fn. 21) makes the suggestive claim that Berkeley, in NTV 12 (quoted above), may have endorsed only a weaker, dispositional form of the transparency thesis. On this view, all mental contents are in principle accessible to introspection under the right conditions and with sufficient attention, even if they are not always consciously perceived. Hatfield also cites NTV 51, 66, and 145 to argue that Berkeley allows for the existence of unperceived ideas, referring to ‘transitions [between visible and tangible ideas] of state of mind—something so “swift” that it is “unperceived”’ (Hatfield 2011: 367). However, Berkeley’s point seems rather to be that the imagination is capable of associating or ‘suggesting’ ideas through processes of which we are not explicitly aware. This reading is consistent with the view that the ideas themselves—especially the ‘suggested’ ones, but also the ‘suggesting’ ones—remain accessible to consciousness, even if we fail to attend to the underlying causal mechanisms; in this case, the rapid transition by which one idea ‘suggests’, or gives way, to the next. Berkeley applies this account not only to the habitual shift of attention from signs, such as visual ideas, to what they signify (e.g., certain tangible ideas), but also to the swift association of heard or read words with their meanings (see, for instance, NTV 140; Theory of Vision Vindicated 48).

A similar position might be attributed to Locke. He occasionally refers to involuntary imaginings—for instance, in his remarks on the association of ideas (Essay II.xxxiii) or in reveries, ‘when ideas float in our mind without any reflection or regard of the understanding’ (II.xix.1). As with Berkeley’s view in the New Theory of Vision, Locke may be read as maintaining that only the perception of ideas—rather than the mental processes that produce them—must be conscious to the subject, and even then, perhaps only dispositionally. Yet, as he clarifies in II.xix.3–4, even in the most distracted states short of completely dreamless sleep, the mind continues to experience ideas and retains at least a minimal degree of attention to their succession (see also Müller 2018: 81). Accordingly, while these passages do not mention volitions explicitly, daydreams or reveries could be taken to involve volitions to imagine, with some conscious content always present—even if the volition itself was initiated by a prior mental operation that escaped explicit or self-reflective awareness.²¹

Despite all these complex interpretative questions I cannot settle here, Baxter could still press the point that if Locke and Berkeley take the mental acts responsible for our dreams to be volitions, then it is uncontroversial that these processes should be accessible to our minds. In the picture of the mind that Locke, Berkeley, and Baxter seem to share with Descartes (albeit with importantly different emphases), volitions are not merely mental processes that can ‘run’ silently in the background, but activities that necessarily involve awareness both that one wills and what one wills.²² These volitions, then, cannot simply be treated as underlying mental mechanisms of which we might not be aware; instead, they are more like the outputs of perceptual processes of which we are necessarily conscious.

Arguably, Locke and Berkeley might stick to their guns and insist that dreams are not produced by volitions, but rather by unconscious and non-voluntary mental operations. However, whether they can genuinely avail themselves of this distinction is debatable. Berkeley, for instance, is quite explicit in identifying acts of the mind with volitions in the first of the Three Dialogues.²³ As should be clear from the above discussion, Locke is hardly any more forthcoming than Berkeley about how unconscious mental operations might fit into his philosophy of mind. Like Berkeley, he explicitly—at least—acknowledges no mental operations apart from essentially conscious volitions and perceptions, and no mental faculties beyond understanding and will (see Essay II.vi; II.xxi.5–6). Relatedly, he does not countenance mental operations of which we have no ideas derived from reflection. This, at any rate, is how Baxter reads him.

Moreover, as Baxter repeatedly points out (e.g., ED: 45), there is something conceptually troubling about the idea that the same mental content can be transparently perceived and yet entirely unconsciously generated by the same mind. In any case, the deeper problem concerns what we might call the infallibility of (self-)consciousness, or what we might call the ‘transparency of perception’: the assumption that if an idea appears to be P (e.g., passive) to a mind, then it really is P. Simply put, ideas—whether waking or dreaming—possess the qualities, and only the qualities, that they appear to have in consciousness. This is particularly problematic for Berkeley, who, as we have seen, does not wish to deny that dream ideas are experienced as passive.²⁴

To make matters worse (as if things weren’t already difficult for Berkeley), abandoning this principle would be an even greater challenge than modifying his commitment to mental transparency, as discussed before. Attempting to preserve his (presumed) understanding of dreaming as an unconscious, non-voluntary mental activity by relinquishing the transparency of perception would undermine both his causal condition of reality and his related argument for God as the cause of our perceptions constituting sensible reality (see PHK 25–26, 29). Since we experience our waking ideas as passive, this transparency thesis allows us to conclude that they must indeed be so, and hence that there must be an external cause producing them. Given that inert matter, like ideas of sense, is passive and already excluded as a cause, we can fairly safely conclude that God is the producer of these ideas (in regular patterns), and therefore these ideas are part of reality (as opposed to unreal or illusory states that we ourselves produce or imagine). However, Berkeley faces serious difficulties defending this crucial argument—and this crucial point about what makes ideas of sense ‘real’—if he allows for the possibility that dreams feel passive but are nevertheless caused by our own minds.

Indeed, our epistemic ability to derive metaphysical truths from the phenomenological passivity of our ideas of sense is central to Berkeley’s anti-sceptical project, as well as to his argument for the existence of God, which rests firmly on our capacity to determine what is real. If, as Baxter argues rather persuasively, our self-consciousness can deceive us in dreams by presenting as passively received an idea that is actually created by our own minds, what guarantees that we do not commit the same error in waking life? Perhaps, as with our dream experiences, we mistake an idea of the imagination created unconsciously by ourselves for a God-given idea of sense. The mere phenomenological fact that it appears otherwise, as Berkeley is compelled to acknowledge, provides no infallible evidence in this regard. From this perspective, it is not surprising that Baxter accuses the theory of dreaming that Berkeley—and, to some extent, Locke—seem to accept of opening the door to solipsism (see ED: 19–20, and note 17). Accordingly, Baxter’s overarching charge that Berkeley is effectively a global dream sceptic also appears to have more force than it might initially seem.

IV. Baxter(ian Principles) to the Rescue

Interestingly, revealing the philosophical depth of his thought, Baxter’s conceptually rich theory not only exposes this problem but might also offer some solutions. For example, had Berkeley adopted Baxter’s spiritual account of the origins of dreams, he could have said—respecting the transparency of perception and the passive phenomenology of dream experiences—that dream ideas are indeed produced by beings other than ourselves, namely, ghosts. In doing so, Berkeley would have avoided having to choose between two options equally untenable within his system: appealing either to unconscious mental activity or to divine intervention to explain the origins of dreams. In other words, with the help of such a theory, Berkeley could both take our phenomenology of dreams seriously and deny the possibility of unconscious or involuntary mental activity, without invoking God as the cause of our dream experiences and thus making them part of reality.²⁵

Unfortunately, the problem of how to accommodate dreaming within Berkeley’s system cannot be fully solved so easily. While this Baxterian proposal preserves the infallibility of self-consciousness—maintaining that what appears passive is indeed passive, caused by beings other than ourselves—it still fails to uphold Berkeley’s argument for God and, more broadly, his distinction between passivity and activity as both a metaphysical/causal criterion and a phenomenological marker of reality. Surely, just because dreams feel passive due to their being actually caused by ghosts, it does not follow that they are real in the same way as our ideas of sense, generated by God. Berkeley might address this concern by supplementing Baxter’s ‘spiritual origin’ theory with the claim that passivity is not a sufficient but only a necessary condition of reality. Accordingly, we infer the existence of God—and the reality of our ideas of sense—not solely from the passivity of our ideas, but also from their remarkable coherence, law-like regularity, intersubjectivity, and usefulness (see DHP: 235; PHK 30–34, 84, which, while not mentioning dreams, might be used to support this broader explanatory framework). This is a common interpretative strategy in the scholarship, exemplified, for instance, by Downing (2021, §3.2.1).²⁶

However, one problem with this strategy is that these additional criteria work much less well in the case of dreams than in that of imagination, since dreams can, at least in theory and sometimes in fact, be profoundly deceptive: not only offering vivid and passive experiences arranged in an internally coherent, regular, and useful order, but also, on occasion, aligning externally with our waking lives. So, if passivity is not regarded as a sufficient phenomenological criterion, one might also question whether these additional features are even jointly sufficient to demarcate reality from dreaming. Indeed, despite Berkeley’s rather nonchalant appeal to them as obvious marks distinguishing real ideas from those of imagination, his remarks in PHK 18 (as discussed above) suggest a not insignificant degree of hesitation about the adequacy of these criteria in relation to dreams. Notably, on the Baxterian proposal currently under consideration, dreams may in fact be even more likely to appear realistic, since they are caused by powerful spirits—beings arguably far more capable of producing internally and externally coherent experiences than our own minds, with their more limited capacities.

More importantly, as noted above, the passivity criterion appears fundamental to Berkeley’s commonsensical conception of reality, serving as a categorical marker—both phenomenological and metaphysical. Unlike vividity, coherence, or regularity, passivity offers an intrinsic and immediate sign of reality: one that does not depend on comparison with, or the sequence of, other ideas, and is therefore not subject to context-sensitive, gradable, or otherwise debatable judgement. In this sense, passivity functions as a better and more reliable guide to the reality of our experiences. Passivity also appears to account for why our sensory ideas exhibit these other indicators: ideas of sense are more vivid, coherent, intersubjective, and law-governed precisely because they are caused by a being like God, who—unlike us—is capable of producing such qualitative and relational features.²⁷

However, the proposed Berkeleyan solution reduces passivity to the status of a mere necessary condition of reality. In other words, while Berkeley may preserve the transparency and authority of self-consciousness even in dreaming, he can do so only at the expense of abandoning felt passivity as a reliable—let alone the surest—phenomenological marker of reality. Worse still, this move would compel him to relinquish the more fundamental metaphysical claim that causal passivity is essential to what makes an experience real, or part of reality, in the first place. Berkeley is unambiguous about its importance. As he writes in PHK 33, ‘the ideas imprinted on the senses by the Author of Nature are called real things’, or in PHK 34, indicating again that ‘the distinction between realities and chimeras’ lies in the fact that ‘real things in opposition to chimeras, [are not] ideas of our own framing’.

But there is perhaps a more promising way to defend Berkeley. Turning Baxter’s own weapons against him, Berkeley might adopt the distinction between two senses of passivity: the mere absence of felt activity (call it ‘non-voluntariness’) and the stronger, positive sense of being acted upon by an external power (‘involuntariness’). Crucially, this distinction could support Berkeley without requiring commitment to Baxter’s peculiar theory of the spiritual origin of dreams. On this view, in contrast to Baxter, Berkeley might concede that in dreams we are passive in the former sense: we do not consciously will or control the ideas that appear. However, upon reflection (available only after waking), we may realise that, while dreaming, we are not passive in the latter sense. Unlike waking perceptions, dream ideas are not caused by an external agent such as God or a spirit. Rather, in dreaming, the mind unconsciously generates its own ideas in a non-voluntary manner, and we overlook this simply because there is no felt sense of imposition, either internally or externally.

If this is right, then Berkeley can maintain the transparency and veridicality of self-consciousness in both dreaming and waking. The passivity of dream experience involves a kind of phenomenological ‘silence’—not, strictly speaking, an error—regarding the cause of our ideas. Self-consciousness does not misrepresent the nature or origin of dream experiences: the mind is indeed not consciously active in voluntarily producing them; it simply fails to register that the ideas are internally caused and does not (mis)represent them either as voluntary products or as externally imposed. In other words, whereas the phenomenology of involuntariness—the feeling of being imposed upon—would point to an external causal source, the phenomenology of mere non-voluntariness, even taken as fully authoritative, is compatible with the mind itself being the (unconscious) source of those ideas. Dreaming, then, reveals certain important limits to mental transparency—and requires, after all, that Berkeley accepts subconscious mental operations—but it does not undermine the authority of self-consciousness in the radical way Baxter thinks, where even solipsism seems to follow. Put simply: the dreamer is not mistaken about, but merely unaware of, the source of their ideas.

This sense of mere non-voluntariness in dreams stands in contrast not only with imagination, where we clearly experience voluntary activity, but also with waking perception, where we not only lack a sense of the source, but positively feel our ideas as imposed on us by a more powerful will. If this suggestion holds, to solve the Baxterian challenge, Berkeley needs only reject Baxter’s claim that dreams, like waking reality, are experienced as positively passive—imposed on us by another mind. One potential worry with this strategy is that it may edge Berkeley closer to Locke’s radical scepticism about the realistic phenomenology of dreams—especially in light of Berkeley’s own remarks that dreams are generally ‘dim, irregular, and confused’, even if sometimes ‘lively and natural’ (DHP: 235). Still, on this view, Berkeley preserves a fundamental phenomenological distinction between imagination, dreaming, and waking perception: (i) imagination involves conscious and voluntary activity; (ii) dreaming involves unconscious and non-voluntary activity; and (iii) waking perception involves involuntary experiences, imposed by another will. Accordingly, he can still account for the subjective indistinguishability of dreaming and waking, even though only the latter involves a stronger sense of imposed passivity. No dreamer, he might argue, is in a sufficiently reflective state to notice this distinction—just as no dreamer can reasonably be expected to recognise inconsistencies between dream content and waking life prior to waking. If so, Berkeley can continue to treat this stronger form of passivity—involuntariness—as both a sufficient phenomenological and metaphysical criterion of reality, and as a valid basis for inferring the existence of God as its cause.

V. Conclusion

As I have tried to show in this paper, Baxter developed a rich and fascinating conception of dreams. While the operation underlying the production of dream images draws on traces left in the brain by past perceptions and retained in memory, the dream world—shaped by the immense creativity of external spirits and our interaction with them—far surpasses anything available to our (waking) imagination. Indeed, dreaming is not parasitic upon perception, nor merely derivative of it: for Baxter, dreams have their own epistemological significance, revealing aspects of our (disembodied) cognitive capacities that waking perception tends to conceal. Moreover, Baxter not only offers a fine-grained faculty psychology by classifying dreams as categorically distinct from our ‘waking fancy’, or imagination—categories that Hobbes, Locke, and Berkeley tended to conflate—but also, by insisting on the sui generis and irreducible nature of dreams, exposes serious blind spots in the philosophy of his predecessors (especially Locke and, by extension, Berkeley), by challenging their accounts of the causes of dreams and their broader assumptions about the nature of the human mind.

I have also sought to show that the Baxter–Berkeley connection on dreams is particularly complex. Baxter not only exposes Berkeley’s difficulty in integrating dreams into his system—a difficulty that, like Locke, Berkeley appears to have largely brushed aside—but, ironically, also provides him with conceptual resources for addressing it. Whether one accepts the Berkeleyan defence I have sketched depends, in part, on whether one agrees with Baxter’s criticism that Berkeley takes the dream argument—the subjective indistinguishability of dreams from waking reality—too far, and thereby undermines the crucial distinction between the unreality of dreams and the reality of sensory ideas. But irrespective of who emerges victorious in this debate, its complexity underscores just how rich Baxter’s thinking was in all matters relating to dreams.

Acknowledgments

I would like to thank Ruben Noorloos and the anonymous reviewers of the Journal of Modern Philosophy for their instructive comments on this paper.

Competing Interests

The author has no competing interests to declare.