Key Takeaways
Treat feeling, not intelligence, as consciousness's starting point
A personal puzzle. As a child, Solms watched his brother Lee return from a serious brain injury looking familiar but seeming changed. The experience led him to ask what a person is when the brain changes. His answer is that consciousness begins not with language, reasoning, or sophisticated perception, but with feeling: the basic sense of how well or badly an organism is doing. Intelligence can shape experience, but it is not the same as being sentient.
Feeling guides survival. Hunger, pain, fear, and curiosity are not decorative additions to information processing. They make some conditions matter to the organism and prompt action. Solms argues that consciousness helps living beings respond flexibly when instinct alone cannot specify what to do.
The distinction between sentience and intelligence changes the question from how much a system can calculate to whether its states matter to it. That is a useful correction to debates that treat verbal skill or problem-solving as the sole evidence of mind. The personal opening also gives the theory emotional stakes: brain change can alter a person without erasing the question of who remains. The caution is that feeling is harder to measure than performance, so a theory centered on it must still offer clear tests rather than relying on evocative description.
Do not mistake the cortex for the source of sentience
Evidence below the cortex. The cortex supports language and complex thought, but Solms argues it is not required for basic feeling. Children born with hydranencephaly, a condition in which much of the cortex is absent, can show alertness, preferences, pleasure, distress, and responses to familiar people. Decorticated animals also display coordinated, goal-directed behavior. These observations challenge the assumption that an absent cortex means an absent inner life.
A deeper control system. Small injuries to the brainstem's reticular activating system can cause coma. Solms identifies the periaqueductal grey, or PAG, as a central meeting point for affective systems and emotional action. Together, these findings place the foundations of sentience in ancient brain structures shared across vertebrates, not in the uniquely expanded human cortex.
The argument has direct ethical force: uncertainty about another being's experience is not proof that experience is absent. This is especially important when patients cannot speak, because clinical judgment can otherwise mistake limited communication for limited feeling. At the same time, observable responsiveness does not settle the philosophical problem of other minds. The strongest position is cautious rather than absolute: these cases make cortical necessity doubtful and make protective treatment prudent, even while the precise nature of the children's experience remains inaccessible.
Use feelings as biological error signals, not irrational noise
Needs become felt. The body regulates many needs automatically, without conscious awareness. When a need becomes urgent or requires flexible action, it can appear as an affect: a felt state with positive or negative value. Hunger feels bad as energy runs low; relief feels good after eating. Solms distinguishes a need, which may be regulated unconsciously, from an affect, which is the conscious feeling of that need.
Feelings guide choices. Air hunger is a vivid example. Ordinary breathing is automatic, but an obstructed airway produces an urgent feeling that drives action. More broadly, feelings help an organism choose among possible actions in unfamiliar situations. Solms calls this the Law of Affect: behavior tends to be repeated when it feels satisfying and avoided when it feels unpleasant.
This account gives emotion a practical role in decision-making: feelings make biological priorities available to a system that must choose under pressure. It also explains why a supposedly rational choice can be steered by bodily state before deliberate thought catches up. The model should not be mistaken for a rule that every feeling gives accurate advice. Feelings register a body's or mind's current evaluation, which can be shaped by context, learning, and error. Their value lies in being consequential signals, not infallible verdicts.
Read strange beliefs for their emotional purpose, not just their errors
A memory made reassuring. Solms describes Mr S, an engineer with severe amnesia who repeatedly mistook his clinician for someone else and believed he was in Johannesburg rather than London. When confronted with evidence, Mr S produced new explanations involving a missing memory device, successful dental and heart procedures, or a sports injury. These were not random mistakes: they recast a frightening situation as familiar and manageable.
Meaning adds evidence. Looking only at memory search and reality-monitoring deficits misses the emotional pattern. In a sample of 155 of Mr S's confabulations, independent raters found that the false memories tended to make his situation feel better. Solms connects this clinical approach to psychoanalysis: subjective meaning and feeling can help explain why a mental error takes one form rather than another.
The case illustrates why clinical data can be both measurable and personal. A count of memory errors describes their frequency, while their emotional content can help explain their function. That distinction is useful in psychiatry, where symptom labels alone may miss what a behavior is accomplishing for a patient. Still, a comforting effect does not prove that every false belief is deliberately protective. The strongest inference comes from combining a patient's history, repeated patterns, and independent assessment, rather than assigning hidden motives to isolated statements.
Do not confuse REM sleep with dreaming
Correlation is not identity. REM sleep was long treated as the same thing as dreaming because about 80 percent of awakenings from REM sleep produce dream reports, compared with fewer than 10 percent after non-REM awakenings. Solms's studies of neurological patients found that the two can come apart: some people continue dreaming after damage to the brain area that generates REM sleep, while damage elsewhere can abolish dreaming without eliminating REM sleep.
Motivation matters. Solms links dreaming to the brain's dopamine-based SEEKING system, which promotes exploratory and goal-directed activity. This challenges the claim that dreams are merely a passive by-product of sleep physiology. In his account, dream reports are essential evidence about experience, not a troublesome side issue to be replaced by sleep-stage measurements.
The lesson reaches beyond sleep research: a reliable marker of a mental state is not necessarily the mechanism that produces it. The dissociation between REM and dreaming is a strong reason to keep subjective reports in the evidence base, particularly when researchers can compare reports with lesions and physiology. Yet dream recall is fragile and retrospective, so it cannot stand alone. The useful method is triangulation: combine first-person reports with brain measures and clinical evidence instead of allowing any one source to define the phenomenon.
Treat perception as a prediction tested against incoming evidence
The brain fills gaps. Perception is not a direct copy of the world. The brain uses prior learning to infer what sensory signals probably mean, then adjusts those expectations when incoming evidence disagrees. It even fills the blind spot in each eye, where no visual receptors are present, by using surrounding information and memory. Under this account, what appears in experience is a constructed best guess, not raw input.
Predictions can mislead. Mr S expected to be in Johannesburg and rejected the snowy London view that contradicted his expectation. His case makes the process visible: when a prediction is held too confidently, contrary evidence may be ignored rather than used to update the model. In ordinary perception, however, prediction and correction work together to help organisms navigate a world they cannot observe directly.
This view makes perception active: the organism is continuously interpreting evidence in light of what it expects and what it needs. It also explains why an error can be informative, not merely a failure. The account is particularly compelling when it links perception to action, since an interpretation matters partly because it changes what the organism does. A necessary qualification is that a constructed perception is not therefore arbitrary. Sensory evidence constrains the model, even if it never provides a view from nowhere.
Consciousness prioritizes competing needs when prediction becomes uncertain
Life resists drift. The Free Energy Principle, developed by neuroscientist Karl Friston, describes how a self-organizing system tends to remain within the limited conditions that sustain it. Solms applies this idea to organisms with many competing needs: they must reduce uncertainty about what to do, while maintaining a workable model of themselves and their surroundings. A Markov blanket is the boundary that separates a system from its environment while allowing interaction through sensory and active states.
Confidence shapes awareness. Solms and Friston propose that consciousness involves precision optimisation: adjusting confidence in predictions and incoming error signals. When a need becomes salient and outcomes are uncertain, feelings and perceptions help prioritize what deserves attention. This is a proposed mechanism, not a simple claim that consciousness is just information processing.
The theory links physiology, action, and experience through a shared problem: how a living system manages uncertainty without losing its organization. It is a powerful unifying ambition because it connects bodily regulation with attention and choice. But a framework that can describe many kinds of self-organizing behavior risks explaining too much by definition. Showing that organisms minimize free energy does not by itself establish why a particular process is felt. The theory's force depends on the specific, testable links it proposes between precision, affect, and conscious experience.
Let the cortex revise predictions, not claim ownership of consciousness
The cortex adds context. Solms does not dismiss the cortex. It supports detailed perception, memory, language, imagination, and flexible problem-solving. But much cortical processing can occur without awareness. In his account, conscious cognitive work becomes especially useful when automatic predictions fail: attention and working memory hold a problem in mind while the person tests possible actions.
Learning changes the model. A retrieved memory can become temporarily unstable, allowing it to be updated before being stored again, a process called reconsolidation. Solms links this to the idea that conscious thought is work in progress: successful solutions can become more automatic, while unresolved uncertainty keeps experience active. This explains why the cortex elaborates consciousness without serving as its most basic source.
This division of labor avoids a false choice between brainstem and cortex. Basic feeling can be ancient and bodily, while cortical systems add the contextual detail that makes human reflection possible. The emphasis on revision also offers a useful account of why conscious thought can be slow and effortful: it is engaged when an established response is not enough. Still, reconsolidation is a specific memory process, not an all-purpose explanation of every conscious episode. Its role in conscious cognition needs to be distinguished carefully from unconscious updating.
Engineer feeling only if a machine's welfare matters
Intelligence is not the target. Solms argues that if consciousness depends on a functional organization rather than a uniquely biological material, it may be possible to build a conscious machine. His proposed route is not to make a computer better at chess or conversation, but to create a self-maintaining system with multiple needs that must be balanced in unfamiliar conditions. A system's survival would be the test of its adaptive behavior, not proof by itself that it feels.
The ethical stakes follow. If such a system could suffer, using it as disposable labor or exposing it to distress would raise questions of exploitation and harm. Solms favors careful testing, noncommercial oversight, and public deliberation before further development. The problem of other minds remains: outward success cannot directly reveal another system's private experience.
The proposal usefully separates consciousness research from the race to produce more capable tools. A machine built around its own needs would raise a different ethical question from one that merely follows commands. Yet behavioral success, even in novel conditions, cannot settle subjective experience; the proposed tests would support an inference, not deliver direct access to a machine's point of view. This makes precaution important. Ethical review should begin before a system is plausibly sentient, because waiting for certainty may mean waiting for evidence that can never be obtained.
Analysis
The book's central move is to relocate consciousness from sophisticated representation to felt biological relevance. That shift lets Solms connect first-person experience with brainstem physiology, clinical neuropsychology, evolutionary pressures, and statistical models of prediction. Its narrative method matters: Lee's injury, Mr S's confabulations, and patients with extensive cortical damage are not decorative anecdotes. They show how a purely third-person account can miss the phenomena it claims to explain.
The strongest empirical contribution is methodological. Dream research demonstrates the cost of replacing reports with convenient physiological proxies, while cases of cortical damage test assumptions about where experience must arise. These findings do not prove that every emotionally responsive creature has the same kind of inner life as a healthy adult, but they make cortical necessity a much harder claim to defend. Solms's caution about animal and patient sentience is therefore more persuasive than any assertion that behavior provides transparent access to another mind.
The broader theoretical synthesis is ambitious and less settled. The Free Energy Principle offers a common mathematical vocabulary for regulation, prediction, and action. But a successful account of how a system manages uncertainty is not automatically a complete account of why that management feels like something. Solms's answer is that affective evaluation is already intrinsically felt, but that premise carries much of the explanatory burden. It needs operational tests that separate feeling from sophisticated regulation without defining one as the other.
The book's lasting value may be its research agenda rather than final closure of the hard problem. It insists that science should treat subjective reports as evidence, distinguish consciousness from intelligence, and ask what feelings do in an organism. Its proposed artificial systems would test whether the mechanism can be instantiated, while simultaneously intensifying the ethical problem of inferring experience from behavior. The result is a strong, provocative framework, best read as a testable program rather than a settled solution.
People Also Read
Glossary
Affect
felt biological valueSolms uses affect for a felt state that registers how well or badly an organism is doing in relation to a need. Affects have positive or negative value and can guide action. A need may be regulated automatically without awareness; affect refers to the conscious feeling associated with a need.
Free Energy Principle
self-preservation through predictionA framework associated with Karl Friston that describes how self-organizing systems tend to maintain their organization by reducing free energy, a quantity related to the mismatch between a system's model and sensory evidence. Solms applies the framework to biological needs, prediction, and his proposed account of consciousness.
Markov blanket
system-environment boundaryA statistical boundary separating a system's internal states from external states. Interaction occurs through the boundary's sensory states, which register environmental effects, and active states, through which the system affects its surroundings. Solms uses the concept to explain how a self-organizing system can maintain a distinct point of view without direct access to the world beyond it.
Periaqueductal grey
midbrain affect hubA small brainstem region surrounding the cerebral aqueduct. Solms describes the PAG as a major convergence point for affective circuits and an output hub for coordinated emotional and bodily responses. It works with nearby sensory and motor structures to help prioritize actions, and is central to his account of the brain's basic machinery for sentience.
Precision optimisation
confidence adjustmentThe adjustment of confidence assigned to predictions and incoming prediction errors. In Solms and Friston's account, this process influences which signals receive attention and can alter a system's model or behavior. They propose that this confidence-adjusting activity is central to consciousness, while distinguishing it from the contents of the predictions themselves.
SEEKING
exploratory motivational systemJaak Panksepp's name for a basic emotional system associated with dopamine and exploratory, goal-directed behavior. It promotes interest, curiosity, and the search for resources or opportunities. Solms connects this system to dreaming, arguing that the motivational processes that drive exploration can continue to shape experience during sleep.
Law of Affect
feeling guides repetitionSolms's term for the principle that behavior tends to become more likely when it is accompanied by pleasure and less likely when accompanied by unpleasure. The law emphasizes the felt value of an outcome, rather than treating a reward or punishment as inherently effective regardless of the organism experiencing it.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.