Consciousness and the Unconscious Mind

Lecture Supplement


So far in this course, we've focused mostly on the principles that govern conscious mental life.  Conscious sensation and perception, conscious learning and memory, conscious thinking and communicating.  Conscious feeling, conscious desires. And that leads us to raise two additional questions:

  1. What is the nature of consciousness?  What causes us to be consciously aware of our percepts, memories, and thoughts, our feelings and desires?  And what good is consciousness so far as behavior is concerned?
  2. And are there aspects of mental life of which we're not consciously aware.  What does it mean, does it mean anything at all, to say we have unconscious sensations, or memories, thoughts, feelings, and desires?

First things first: let's talk about consciousness.

An Introduction to Consciousness

When we try to understand mental states such as percepts, memories, thoughts, feelings, and desires, we are usually trying to understand conscious mental states -- that is, percepts, memories, etc. of which we are consciously aware. 

  • In his famous statement, Cogito, ergo sum ("I think, therefore I exist"), Rene Descartes (1637) shifted the concerns of Enlightenment philosopher from metaphysics and ethics to epistemology, and began to lay the groundwork for the later emergence of scientific psychology.  It was his conscious awareness of his thought processes that led to his insight. 
    • Descartes' philosophical position is known as dualism, because he construed minds and bodies are two different kinds of things.
    • Psychology, as a science, is inherently dualistic in nature.  Al psychologists recognize that the brain is the physical basis of mind, but from its 19th-century beginnings psychology has analyzed mental life without regard to its biological basis.  For example, the principles of perception and memory, reasoning and problem-solving, discussed earlier in this course were discovered by looking at behavior (chiefly in the form of self-reports and reaction-time), not brains, and inferring the underlying mental, not biological, states and processes. 
  • William James, began his Principles of Psychology (1890) by defining psychology as "the science of mental life" -- by which he meant conscious mental life.
    • And just to make it clear, when he published his abridged Psychology: briefer Course (18929), he defined psychology as"the description and explanation of states of consciousness".

Actually, however, until recently psychology has not had much to say about consciousness itself.  That wasn't always the case, however.  

  • As noted in the Introduction, scientific psychology began, in the early 19th century, expressly as the study of consciousness, with the development of psychophysical laws relating physical stimulation to conscious sensation. 
  • And later on in the 19th century, the structuralists attempted through introspection to identify the basic building blocks out of which conscious mental states were assembled.
  • But concerns about introspection as a method, and constant debates among the Structuralists about which were the basic elements of consciousness proved unresolvable, so the Behaviorists decided simply to abandon consciousness, and mental life in general, as the subject matter for psychology.
  • Still, interest in consciousness persisted among some theorists, as seen most clearly in the Gestalt approach to perception, and research on the span of apprehension, or the number of items that could be kept in mind at once.
  • It was not until the Cognitive Revolution, beginning in the 1960s, that psychologists began to broach the topic of consciousness again, in the form of studies of attention, sort-term memory, and mental imagery. And even then, their approach was mostly piecemeal -- that is, they assumed that once they understood enough about perception, or memory, or problem-solving, then an understanding of consciousness would naturally emerge. And there were also contradictory trends even within cognitive psychology and cognitive science -- particularly the computer metaphor of the mind, which implied -- since after all computers themselves were not conscious entities -- that cognition might be understood without reference to consciousness after all!

It is only very recently that psychology, and the other cognitive sciences, began to try to analyze consciousness directly.

The goal of all this research and theory is to take a truly scientific approach to consciousness, leading to a public, objective description and explanation of private, subjective conscious experiences.  That seeming contradiction -- between what's public and what's private, what's objective and what's subjective -- makes consciousness research so difficult, and so exciting.

The scientific study of consciousness attacks four basic problems.

  • The Mind-Body Problem
  • Knowledge of Mind
  • Altered States of Consciousness
  • Unconscious Mental Life

Mind and Body

First and foremost is what philosophers since Descartes have called the mind-body problem, which asks about the relations between bodies and minds.

  • In modern science, the traditional mind-body problem has been resolved in favor of materialism -- that is, that there are only physical substances, and so the brain is the physical basis of mind. 
    • Materialism has become the default position for the study of consciousness, and gives rise to what the philosopher David Chalmers has called "the hard problem" of consciousness:
      • How do neural states give rise to conscious states?
      • And why doesn't mental life just all go on "in the dark"?
  • Those are pretty difficult questions, so most consciousness researchers have contented themselves with identifying the neural correlates of consciousness -- those aspects of neuroanatomy and neurophysiology that are associated with conscious mental states.
    • Of course, the neural correlates of consciousness will depend on how consciousness is defined, so adopting the materialist stance doesn't mean you can abandon philosophy or psychology.
  • The method in these studies is usually to experimentally manipulate some aspect of consciousness, and then examine how the brain changes.
    • Induce an altered state of consciousness, as with drugs such as LSD.
    • Examine the neural differences between explicit and implicit memory and other explicit-implicit dissociations.

Identifying the neural correlates of consciousness, and working on a solution to the "hard problem" of consciousness, would seem to be enough.  But it turns out that there are three more aspects of the mind-body problem.

Mind-Body interactions.  Descartes thought that mind and body were composed of separate kinds of substances, material body and immaterial mind, which made him a dualist.  But he also understood that mind and body interacted, which means he was an interactive dualist.  To paraphrase his own examples:

  • A part of the body, the eye, mediates the conscious experience of seeing.
  • The person's conscious desire to move his arm causes his arm to actually move.

These days, we recognize two more interesting forms of mind-body interactions.

  • First, there is the effect of the mind on the body.
    • Psychosomatic interactions, in which the person's beliefs, attitudes, feelings, or desires affect the physiology of his body.
    • The most widely discussed psychosomatic interactions are placebo effects -- where, for example, the person's belief that he is taking an analgesic drug leads to significant pain relief.
  • Second, there is the effect of the body on the mind.
    • Embodied cognition, in which the physical state of the body affects some aspect of perception, memory, or thinking.
      • For example, a hill looks steeper if you're climbing it with a heavy pack on your back.
    • Interactions between emotion and cognition and motivation and cognition may be of this sort.
      • For example, being hungry might make servings of food look bigger than they are.

Body Without Mind.  This is the question of how important mental activity is to behavior.  Can we get along without any conscious mental states at all?

  • In their thought experiments, philosophers often discuss the example of zombies, which are creatures which have no consciousness, and operate solely on the basis of reflex, taxis, and instinct.
  • At the same time, cognitive psychology acknowledges a difference between automatic and controlled processes, where automatic processes, essentially, operate in a reflex-like fashion, independent of conscious awareness and conscious control. 
  • Automaticity, in turn, raises the problem of free will, and how much of it we actually have.  And some social psychologists have gone so far as to suggest that human behavior is so dominated by automatic processes that we don't have any free will at all -- that free will is an illusion.

Mind Without Body.  Psychologists believe, almost as an article of faith, that the brain is the physical basis of mind, and that when the brain stops working, the mind does too.  This is what is known as the materialist stance.  But there are some researchers who believe that mental states can exist independent of physical states like neural activity.

  • Parapsychology studies the possibility that thoughts can be transmitted from one mind to another in the absence of any physical medium -- as when mediums claim to be able to see the future or read other people's minds.
  • 19th-century spiritualism was based on the belief that people could communicate with the dead.
  • In the 20th and 21st centuries, spiritualism was revived with an interest in out-of-body experiences and near-death experiences.

Introspection and Knowledge of Mind

Descartes came to his "cogito" through a process of introspection -- by "looking into" his own mind, and discovering that there were sensations, thoughts, feelings, and the like.  That's how we all know that we're conscious, and introspection has been the primary method by which philosophers (and psychologists, too, for that matter) have tried to understand consciousness.  And, for that matter, this is how we know what's on other people's minds -- by asking them what they're thinking, or feeling, and taking their replies seriously.  But this raises the question of how well we know our own minds.  Maybe we really don't know what we're feeling, or maybe our conscious thoughts don't really influence our behavior, because all the action really lies with unconscious mental states and processes -- things running through our heads that determine our behavior, of which we are not consciously aware and over which we have no conscious control.  Or maybe our experience, thought, and action is determined by some combination of conscious and unconscious mental processes.

So one question has to do with the limits on introspection

  • Just how accurate are our introspections?  How well can we know our own minds?
  • How accurate are other people's reports of their introspections?  How much can we trust what they say about themselves?
  • What can we say about consciousness is creatures that can't speak to us -- infants, nonhuman animals, even alien beings if we ever encountered them?
  • How would we know if a computer or robot had consciousness?

This problem is particularly critical when we ask questions about the origins of consciousness.

  • Ontogenetically speaking, at what point can we consider the developing child to be conscious?
    • Are infants conscious, the same way adults are?  And how would be know, given that they can't report their introspections to us?
    • And if we're going to talk about infant consciousness, we might just as well raise the question of whether, and when, and to what extent, an unborn fetus has anything like consciousness.
  • Phylogenetically speaking, at what point did consciousness evolve? This is known as the problem of animal awareness.
    • Do primates, like chimpanzees, have anything like consciousness?  If so, what kind of consciousness do they have?
    • And what about other mammals, like dogs and cats?  Descartes denied that nonhuman animals have consciousness, arguing that they were just reflex machines.  Was he right?
    • And if we're willing to grant consciousness to some mammals, how about other vertebrates?
      • How about invertebrates?
      • And, for that matter, what about plants?

Again, basically, the question is: How would we know?  Since these creatures can't talk to us, what behavioral or physiological criteria could we use to analyze their mental states?

For the most part, research has focused on two paradigms:

  • Mirror self-recognition, in which we determine whether a subject -- human, infant, chimpanzee, elephant, whatever -- can recognize itself in a mirror.  Mirror self-recognition is thought to require a degree of self-awareness -- of what one looks like, and what one is doing -- that is one of the hallmarks of consciousness.
  • The "false-belief test", in which we determine whether a subject -- again, it doesn't matter what species or stage of development -- can recognize that another creature's mental states -- beliefs, feelings, or desires -- are different from its own.  Again, passing the FBT requires knowledge of one's own mental state, as well as the ability to infer the mental state of someone (or something) else -- and, finally, the ability to compare them. 

Altered States of Consciousness

The combination of self-reports with behavioral and physiological measures is also important in the study of altered states of consciousness (ASCs), whether these are induced by drugs or other techniques, such as hypnosis or meditation.

  • Subjects typically report that they are in an ASC - -that is, that their state of consciousness seems very different from usual.
  • And we can employ behavior and physiological measures to corroborate these introspections.

In this way, a number of ASCs have been studied experimentally:

  • Coma, and the persistent vegetative state (PVS) -- in which brain-damaged patients do not appear to be conscious.
    • Recently, some researchers have argued that at least some comatose or "vegetative" patients might be conscious after all, a condition now known as the minimally conscious state
    • Obviously, behavioral and physiological measures are critical to the task of identifying which patients might be even minimally conscious.
  • Surgical anesthesia also counts as an alteration in consciousness.
    • In general anesthesia, the patient is rendered unconscious while some surgical operation is performed.
    • In conscious sedation, often used in outpatient surgery, the patient is conscious during the procedure, but has little or no conscious recollection afterward.
  • In sleep, we all appear to lose consciousness, at least once a day -- more, if you like to nap!
    • There are various stages of sleep, each of which might count as a separate state of consciousness.
    • Dreams, in which we have hallucinatory experiences, also count as ASCs.
  • In hypnosis, subjects experience alterations in perception and memory in accordance with the suggestions of a hypnotist. 
    • In the classic instance, these alterations are subjectively compelling to the point of delusion.
  • What used to be called hysteria is now separated into the dissociative and conversion disorders.  Whatever you call then, the hallmark of "the hysterias" is that the patient has lost some aspect of conscious awareness or control, in the absence of any demonstrable brain insult, injury, or disease that would account for the patient's symptoms.
    • In the dissociative disorders, the loss of conscious awareness affects memory and identify.
      • In psychogenic amnesia, the patient loses conscious access to some aspect of autobiographical memory.
        • Psychogenic amnesia is also called dissociative amnesia.
      • In psychogenic fugue, the person also loses awareness of his or her identity.
        • Psychogenic fugue is also called dissociative fugue.
      • In multiple personality disorder (MPD), the person shifts back and forth among two or more identities, each with its own associated fund of autobiographical memories.
        • When the patient is in one "ego state", he or she is often amnesic for events that occurred while in the other(s).
        • MPD is also called dissociative identity disorder.
    • The conversion disorders come in two major forms.
      • Psychogenic losses of conscious sensation or perception, such as blindness, deafness, or tactile anesthesia.
      • Psychogenic losses of voluntary motor control, such as paralysis of a limb or aphonia. 
  • In meditation, the practitioner's attention is so tightly focused on some target (such as breathing or a mantra), that he becomes unaware of anything else going on around him.
    • Something similar happens in absorption, where a person is "swept away" by a piece of music, a painting, a sunset, or the like.
    • And the opposite happens in daydreaming, where a person's mind "slips away", and he loses focus on the task at hand.
  • And then, of course, there are the "psychedelic" states induced by mind-altering drugs.
    • "Major" psychedelics such as LSD
    • "Minor" psychedelics such as marijuana.

All of these ASCs can be studied by the methods of experimental psychology, as they have developed since 19th-century psychophysics.

  • And, when combined with neuroscientific techniques such as brain-imaging, might shed important light on the neural substrates of consciousness.

Unconscious Mental Life

And this same combination of self-reports, behavioral measures, and physiological measures is also critical to the question of the unconscious.  The idea that our experience, thought, and action might be influenced by mental states and processes of which we are consciously unaware, and over which we have no conscious control, has been part of popular culture at least since the time of Sigmund Freud (1856-1939) -- and has been rejected by many scientific psychologists for that long, as well.  However, recent research has suggested that the concept of unconscious mental life might have some scientific respectability after all.

First, there is the distinction, discussed in the lectures on Memory, between automatic and controlled mental processes.  Automatic processes occur unintentionally, in a reflex-like fashion, in response to particular stimuli, and often are executed outside of awareness.  For this reason, they are often considered to be unconscious in the strict sense of the term. 

  • Cognitive psychologists have developed a set of behavioral criteria by which we can distinguish between automatic and controlled processes.
  • And they have also developed methods that can help quantify the relative contributions of automatic (unconscious) and controlled (conscious) processes to task performance. 

And then there is the distinction, first demonstrated in amnesic patients, between explicit and implicit memory (this was also discussed in the lectures on Memory).

  • Amnesic patients, by definition, have no conscious recollection of events that occurred since their brain injury.
  • Yes they show priming and other effects that can be attributed to these events.

In this sense, implicit memories are unconscious memories, which influence our experience, thought, and action in the absence of conscious recollection.

Over the past couple of decades, the explicit-distinction has been extended to other aspects of cognition, and even to emotion and motivation.

  • Dissociations between explicit and implicit memory have also been documented in neurologically intact subjects, especially under conditions of degraded encoding.
  • As discussed in the lectures on Perception, similar dissociations between explicit and implicit perception have been documented in the literature on "subliminal perception" -- again, under conditions in which subjects' attention has been distracted from the stimulus.
    • And similar dissociations between explicit and implicit perception have been observed in neurological patients experiencing blindsight, hemispatial neglect, or prosopagnosia.
  • A dissociation between explicit and implicit learning has been documented in neurological patients and normal subjects who have acquired some new knowledge through experience, even though they are not consciously aware of this knowledge.
  • And we can see dissociations between explicit and implicit thought  - -where, for example, a subject has an intuition that a problem is soluble, even though he doesn't actually know what the solution is. 

And the explicit-implicit distinction has even been extended beyond cognition to emotion and motivation.

We'll discuss these trends at greater length in a later section on The Unconscious Mind.  For now, though, let's see how experimental psychology can shed light on some of these problems.

Anesthesia, Coma, and the Minimally Conscious State

In a highly influential article, Laureys (2005) has proposed that there are two components of consciousness:

  • Awareness of oneself and the environment, or the content of conscious experience;
  • Wakefulness. or the level of consciousness.

A wide variety of different states of consciousness can be conceptualized within this framework.

  • At the lowest end of both continua are patients in a coma, who lack both wakefulness and awareness. 
  • And very close to them are surgical patients undergoing general anesthesia, which is sometimes characterized as a "controlled coma". 
  • The persistent vegetative state has been characterized as "wakefulness without awareness", because these patients seem to go through something like the normal sleep cycle.
  • Various stages of sleep fall along the diagonal.
  • Farthest away from coma are people like you and me right now (presumably), fully awake and fully aware of what we are doing and experiencing.

General Anesthesia

Because the anesthesiologist puts the patient "to sleep" and then "wakes him up" again, general anesthesia is sometimes characterized as a "controlled coma".  The first demonstration of general anesthesia, on October 16, 1846, is celebrated in medical schools worldwide as "Ether Day", and is memorialized in this painting on view at the Countway Medical Library at Harvard Medical School.

Modern anesthesia is known as "balanced" anesthesia, because it consists of a "cocktail" of different drugs:

  • A sedative to reduce the patient's anxiety and ease the induction of anesthesia.
  • The anesthetic agent itself, to induce the "controlled coma".
  • A muscle relaxant, to reduce reflexive responses to what are charmingly called "surgical stimuli".
  • An analgesic to reduce post-operative pain.

But how do we know that the patient is really anesthetized?  Put another way, how would we know if an ostensibly anesthetized patient regained consciousness during surgery?  Note that he couldn't tell us: even if the anesthetic didn't take, or wore off, he'd be paralyzed by the muscle relaxant, and unable to communicate (this happens, but it is vanishingly rare).

Until recently, anesthesia was assessed clinically in terms of the patient's behavior during the surgery.

  • Lack of response to verbal commands or "surgical stimulation".
  • No indications of pain during the procedure, and no reports of pain post-operatively.
  • No memory for surgical events.

An alternative method is the PRST score. in which the anesthesiologist looks for various autonomic signs of pain or awareness during the operation:

  • P: Increase in blood pressure.
  • R: Increase in heart rate.
  • S: Sweating
  • T: Secretion of tears.

Another method of monitoring anesthesia makes use of the EEG power spectrum, measured continuously by electroencephalograph (EEG) electrodes attached to the patient's skull.  The EEG signal can be decomposed into a number of "bands":

  • The "waking" EEG consists of a mix of "alpha" and "beta" activity:
    • Alpha, with a frequency of 8-12 cycles per second.
    • Beta, with a frequency of 18-30 cps
  • In sleep, EEG activity slows down considerably:
    • Delta, with a frequency of 0.5-4 cps.
    • Theta, with a frequency of 5-7 cps.  
  • Monitoring the power spectrum, the anesthesiologist tries to keep the EEG wave at a very low frequency:
    • Median frequency = 2-3 cps -- in the "delta" range.
    • Maximum frequency, or "spectral edge" < 8-12 cps, below the "alpha" range.

A very high-tech option is bispectral analysis (BIS), a proprietary device that monitors various aspects of the EEG:

  • The BIS algorithm computes a bispectral index based on such parameters as the amount of high- and low-frequency activity
  • The bispectral index itself is calibrated so that a score of 100 means "awake"
    • A score of 86 reflects a 50% reduction in recall of material presented during anesthesia.
    • A score of 64 reflects a 95% reduction in recall.
    • As a matter of clinical practice, a patient is considered to be adequately anesthetized if his score is < 60.

Many measures of anesthetic depth are based on memory -- that is, the patient's inability to remember surgical events (which, frankly, would seem to be pretty memorable) is taken as evidence that they really were unconscious during surgery.  And so the bispectral index, for example, counts as a physiological correlate of consciousness.  But it turns out that, while anesthesia abolishes conscious recollection of surgical events, some memory may be preserved in the form of implicit memories.

In an experiment in my laboratory, we worked with a group of patients who were undergoing elective surgery under isoflurane anesthesia.  After they were adequately anesthetized, we presented a recording of paired-associates of the form ocean-water or butter-knife.  We then tested memory for these paired associates after the patients were awake in the recovery room, and again on followup 14 days later.

  • Cued recall, an expression of explicit memory, was very poor.  That is to say, when presented with the cue ocean, the patients did not remember that it has been paired with the response water.
  • So was recognition, generally considered to be an even more sensitive index of explicit memory.
  • On a free-association test, however, these same patients showed a significant priming effect.  That is, when presented with the cue ocean, and simply asked to respond with the first word that came to mind, they were more likely to generate the target water than control patients, who did not hear the list.

So, while adequate surgical anesthesia abolishes conscious recollection of surgical events almost by definition, at least some anesthetic cocktails appear to spare implicit memory.  The fact that anesthetized subjects show priming effects, though, does not mean that they were actually conscious during their surgery.  What it means is that some degree of unconscious information processing is possible even in subjects who are adequately anesthetized by all clinical criteria. 

Coma, the Vegetative State, and the Minimally Conscious State

The assessment of consciousness is brought to the fore by coma, which can be characterized as a complete loss of consciousness.  While vegetative functions (e.g., heart rate, respiration, digestion, elimination) are preserved, the patient cannot communicate.  He does not respond to stimulation in any modality, and shows no signs of emotion.  The patient's eyes are permanently closed, but he shows no signs of the sleep cycle.  The severity of coma is clinically measured by the Glasgow Coma Scale, which assigns numbers to the patients "best" ocular, verbal, or motor response to commands.  Scores on the GCS range from 3-15, with a score less than 8 indicating a "severe" coma.  

Comatose patients may recover from the coma.  Or they may move into the persistent vegetative state.  PVS is generally characterized as "wakefulness without consciousness".  The patient is still uncommunicative, but may show partial reflexive response to stimulation -- e.g., auditory or visual startle, or even brief orientation to the stimulus; there are few if any signs of emotion, other than occasional reflexive crying or smiling.  However, PVS patients appear to go through something like the diurnal sleep-wake cycle, at least insofar as their eyes will periodically open or close.  About 10% of patients will move into PVS within the first month after lapsing into a coma; after that, the likelihood declines progressively, to about 1% after 12 months.  About 40% of VS patients will recover some degree of consciousness within 6 months of diagnosis; after that period, the likelihood of recovery is very slim. 

Coma and PVS differ in terms of the site, and perhaps extent, of brain damage.  In coma, the damage appears to be confined to an area of the posterior brain stem known as the reticular formation.  In PVS, the damage tends to be confined to the diencephalon, particularly the thalamus, leaving the RF intact.

Coma and PVS contrast markedly with another condition, known as the locked-in syndrome, in which the patients are clearly conscious, but cannot engage in (much) voluntary motor behavior, including speech and gesture.  The symptoms associated with the locked-in syndrome include anarthria (loss of articulate speech), aphonia (inability to vocalize), and quadriplegia (total paralysis of the limbs).  Still, these patients have preserved auditory and visual function, as indicated by the startle reflex and orienting response; and also by visual localization and fixation, and preserved pursuit eye movements.  The preservation of eye movements also means that "locked-in" patients can communicate with others by means of blinking and other eye movements; they also show signs of preserved emotion.  While coma and PVS result from damage to the reticular formation and/or thalamus, the locked-in syndrome is associated with damage to the anterior portion of the brain stem, particularly the pons.  This explains how patients get "locked in", and also how they are able to communicate via eye movements.

  • Most motor pathways pass through the anterior brainstem.  Therefore, damage to this area effectively prevents motor signals originating in the frontal cortex from reaching the spinal nerves that control the skeletal musculature.
  • But the damage lies at or below the point where the trigeminal nerve (Cranial Nerve V) connects to the brainstem. 
    • Thus, this spares the olfactory (Cranial Nerve I), and optic (II) nerves, permitting afferent signals to reach the brain (the auditory nerve connects directly to the brain through the medial geniculate nucleus, and bypasses the pons entirely).  
    • And it also spares the oculomotor nerve (III) and the trochlear nerve (IV), which carry efferent signals which move the eyes.

The difficulty in determining whether a PVS patient is actually conscious, is vividly illustrated by the case of Terri Schiavo, a young woman who went into a coma in 1990 following respiratory and cardiac arrest, and later lapsed into PVS.  In 1998 her husband petitioned to remove life supports, but her parents opposed the request, on the grounds that Ms. Schiavo showed signs of consciousness.  Eventually, the State of Florida and even the US House of Representatives got embroiled in the dispute.  In 2002, a CAT scan showed very little residual cortical tissue -- a fact confirmed in 2005, when Ms. Schiavo finally died.

In fact, some PVS patients may actually be in a minimally conscious state, in which they show partial, inconsistent evidence of consciousness.  At times, such patients may seem to be communicating with other people, as shown by vocalization that is contingent on environmental events, and even spontaneous verbalization and gesture.  They will also show partial and inconsistent responses to stimulation, such as auditory localization and command-following.  They may engage in sustained fixation on visual objects -- and even, if  inconsistently, engage in visual pursuit.  They will localize noxious stimuli, and engage in automatic responses to them.  In some instances, they can reach for objects, and accommodate their hands to their shape.  They may also display emotional responses such as smiling and crying, contingent on environmental events.  The diagnosis of MCS is crucial, because while we might be willing to withdraw life supports from a patient in a coma or vegetative state, most physicians and families would be reluctant to do so for someone who was even minimally conscious (hence the controversy over Terri Schiavo). 

So how to tell who is who?  One case involved a 23-year-old woman who suffered a traumatic brain injury during an auto accident, and received a diagnosis of PVS.  After 6 months, she was run through an fMRI protocol.  When presented with speech versus noise, she showed some activation in the speech areas of the brain.  When asked to imagine playing tennis, she showed activation in the supplemental motor area (SMA) of the frontal lobe; and when asked to imagine walking around her house, she showed activation in the posterior parietal cortex.  Apparently, she had enough consciousness to make discriminative responses at the level of brain activity -- even though she was physically unresponsive. 

But not all PVS patients are actually MCS. A subsequent study by Monti et al. (2010) tested a larger group of 54 patients, including 23 patients who were diagnosed PVS and 31 patients diagnosed MCS.  When asked to imagine hitting a ball on a tennis ball or walking on a familiar street or in their house, some of these patients showed patterns of activation similar to those shown by 16 healthy controls.  But these were a distinct minority: 4 of the PVS patients, and only 1 of the MCS patients.  [Note: This is surprising, as you'd expect more of the MCS patients to show fMRI evidence of consciousness, compared to the PVS patients.] 

Still, the ability of some patients to respond to differential commands, as evidenced by brain activity, raises the possibility of establishing two-way communication with these patients, much as we can use eye movements with locked-in patients (as illustrated by the book and movie, The Diving Bell and the Butterfly). 

In fact, Monti et al. tried just such a tack, posing factual questions to the 5 patients who showed differential fMRI responses in their experiment.  Only 1 subject responded accurately and reliably to questions such as "Do you have any brothers?" -- that's 1 out of 54, amounting to approximately 2% of the patients studied.  So, this is possible, but one shouldn't depend on it.

The Owen-Monti studies establish that it is possible to establish two-way communication with patients in PVS or MCS -- though they also establish that this happens only infrequently.  However, fMRI is extremely expensive, and for diagnostic purposes it would be useful to have a more efficient technology available.  In a study using EEG (much cheaper and easier than fMRI), Cruse et al. (2011) studied a large group of PVS and MCS patients, asking them to try to squeeze their right hand or wiggle their left toe.  Blind raters, looking at the pattern of EEG activity (in the left and right motor cortex, respectively), were able to classify the patients' commands correctly for 75% of the controls, but only about 20% of the patients.  It's not nothing, but it's not a lot.  [And again, note that there was no difference in rate of response between PVS and MCS patients -- 19% and 22%, respectively.] 

So far, this research indicates that it is possible to assess residual consciousness in coma (broadly defined to include both PVS and MCS as well). In particular, studies of PVS and MCS provide some evidence of intentional, conscious activity, in terms of specific responses to instructions.  But this is true only in small minority of patients.  It is simply not the case that all patients diagnosed with PVS, or even MCS, are actually conscious.  

  • On the basis of the evidence so far, we have to be skeptical of the clinical criteria for MCS.  By any standard, MCS patients should be more responsive than PVS patients in the fMRI and EEG paradigms.  But they're not.
  • Both fMRI and EEG can be employed to make the diagnosis of MCS more rigorous.  Regardless of their clinical diagnosis, only those patients who actually show differential responsiveness on brain-imaging measures should be considered MCS.
  • In those MCS patients in whom reliable two-way communication can be established, a question for medical ethics is whether the paradigm should be used to enroll the patients in making decisions about their own medical care -- for example, the question of whether to limit treatment or even withdraw life supports.

Overcoming Conscious Shyness

In an important essay, the philosopher Owen Flanagan (1992) discussed some reasons why psychologists and cognitive scientists have been reluctant to undertake the study of consciousness. 

  • Positivistic Reserve: Ever since the Behaviorist Revolution, psychologists have focused their research on objective data based on public observation, such as environmental stimuli and behavioral responses to them.  They have been reluctant to base their science on subjective variables that are inherently private -- like subjects' self-reports of their own mental states.
  • Piecemeal Approach: Even those who retained an interest in consciousness generally viewed that an understanding of consciousness would emerge from more specific studies of sensation and perception, memory and thinking, emotion, and the like -- that, at some point, we would know enough about these topics that we could "put it all together" and understand consciousness.
  • Conscious Inessentialism: Still, one of the implications of the computer model of the mind, and the idea of human information processing, was that consciousness might not be particularly important for understanding mind and behavior.  That is, most everything that we do consciously, we could do just as well (or perhaps better) unconsciously.
  • Epiphenomenalist Suspicion: And to take conscious inessentialism to an extreme, it might be that, while we are conscious of what is going on around us, and of what we are doing, that consciousness is essentially an illusion, and has nothing to do with what is really going on.  Put another way, that consciousness plays no causal role in behavior.

The least that can be said, on the basis of the studies reviewed here, is that consciousness can be measured objectively.  We can tell when someone is anesthetized, or in a coma.  And if we suspect that the person is, despite appearances, at least minimally conscious, we have techniques to diagnose even that minimal level of consciousness objectively.  So at the very least, we can begin to abandon our positivistic reserve, and begin to study consciousness directly.  In this way, we'll find out whether consciousness is essential for behavior -- or whether it is an illusion.

Not the lest of these questions is that of unconscious mental life.  That is, whether, and to what extent, our conscious experience, thought, and action is influenced by unconscious percepts, memories, thoughts, feelings, and desires.  In order to address this question scientifically, we have to have objective indices of unconscious as well as of conscious mental life. 

It's to this problem that we now turn.

The Unconscious Mind

Some psychologists have rejected the notion of unconscious mental life out of hand, as a contradiction in terms.  For them, having a mind and being conscious are pretty much one and the same thing.  But others have suggested that there might be more to mental life than that we are consciously aware of.  Writing before the advent of scientific psychology, Immanuel Kant thought that it might very well be possible to have ideas, yet not be conscious of them, and that we could become aware indirectly of these unconscious ideas by observing their effects on our experience, thought, and action.

But first, before we start talking seriously about unconscious mental life, let's be clear on what we mean by consciousness.  Consciousness entails two somewhat different things.  First, monitoring ourselves and our environment, so that percepts and memories, thoughts, feelings, and desires come to be accurately represented in our phenomenal awareness.  Second, controlling ourselves and our environment, so that we are able to voluntarily initiate and terminate behavioral and cognitive activities -- to be able to voluntarily control not just what we do in our behavior, but also what we think, and what we feel, and what we want.  Conscious awareness and conscious control are two different things, but they're obviously related, because it would seem to be difficult, if not impossible, to consciously control something that we're not consciously aware of.

The Unconscious Before Freud

In the last quarter of the 19th century, as the new science of psychology began to emerge from its roots in philosophy and physiology, consciousness was at the center of the enterprise. The early psychophysicists, such as Ernst Weber (1795-1878) and Gustav Fechner (1801-1887) were concerned with the relations between the objective, physical properties of environmental stimuli and the subjective, psychological properties of the conscious mental states to which they gave rise. A major goal of 19th-century psychophysics was to determine the absolute and relative thresholds for conscious experience. These were, respectively, the minimum intensity required for an observer to become aware of the presence of a stimulus, and the minimum increase or decrease in intensity required for an observer to be aware that the stimulus had changed. The methods of psychophysics, such as the "just-noticeable difference", the method of constant stimuli, and the method of adjustment, were the first attempts to quantify subjective experience. The psychophysical laws, which sought to relate every psychological property of a sensation to some physical property of the corresponding experience, were the first attempt to determine the physical basis of consciousness.

Beginning with Wilhelm Wundt (1832-1920) and his American protege E.B. Titchener (1867-1927), the school of psychology known as structuralism attempted a kind of "mental chemistry" which sought to analyze conscious experience into its constituent sensations, images, and feelings (Titchener, 1898). The structuralists' preferred method of investigation, called introspection, assumed that people had accurate conscious awareness of their own mental states (E.G. Boring, 1953). The structuralists had a particular perspective on psychophysics. In contrast to physics, which they defined as the study of the facts of experience independent of the observer, psychology was defined as the study of these same facts as dependent on the experiencing individual (Titchener, 1929/1972). In this way the very first school was, first and foremost, devoted to the study of conscious experience.

Nevertheless, structuralism was hounded by a number of difficulties (E. G. Boring, 1950; Hilgard, 1987). For example, a protracted debate between Titchener and Oswald Kulpe (1862-1917), over whether one could have thoughts without images, revealed that introspection was not always a reliable method. Nevertheless, we owe to the structuralists the discovery of the basic dimensions of certain sensory and emotional experiences. For example, although Sir Isaac Newton (1642-1727) discovered that a prism refracts light into seven primary colors, from a psychological point of view there are only four: red, yellow, green, and blue. This kind of work, in turn, laid the basis for the later search for the neural correlates of what philosophers call sensory qualia. For example, we now know that the physiological basis for color vision is a set of four neural systems, arranged as opponents: red-yellow, and green-blue (DeValois, 1965; Hering, 1878/1964; Hurvich & Jameson, 1957). Without the structuralists' analyses of conscious experience, neuroscientists might well have searched long and in vain for seven neural systems corresponding to the physical primaries. The achievements of structuralism were summarized by E.G. Boring (1933) with a monograph appropriately entitled The Physical Dimensions of Consciousness.

Even William James, opposed as he was to the doctrines of structuralism (Hilgard, 1987), embraced a version of introspection as his preferred research method (he had a collection of brass instruments, but he hated using them). James worried a great deal about the liabilities of introspection, and sought to supplement the technique with experimental methods. James began his Principles of Psychology with the assertion that "Psychology is the science of mental life" (James, 1890/1980, p. 1). By this he meant conscious mental life -- as James made abundantly clear in the Briefer Course (1892/1980, p. 1), where he adopted G.T. Ladd's (1887) definition of psychology as "the description and explanation of states of consciousness as such". For James, as for Descartes, consciousness came first, even before sensation: "The first fact for us then, as psychologists, is that thinking of some sort goes on" (James, 1890/1980, p. xxx).

At the same time, both James and the structuralists understood that there was more to mental life than was accessible to introspection. The notion that unconscious processes are important elements of mental life is commonly ascribed to Sigmund Freud, the founder of psychoanalysis, but in fact it was an old idea before Freud was even born (Ellenberger, 1970; D. B. Klein, 1977; Whyte, 1960).

The Threshold of Consciousness

At the beginning of the 18th century, the German philosopher Gottfried Wilhelm Leibniz (1646-1716) had written that our conscious thoughts are influenced by petites perceptions -- sensory stimuli of which we are not aware:

...at every moment there is in us an infinity of perceptions, unaccompanied by awareness or reflection.... That is why we are never indifferent, even when we appear to be most so.... The choice that we make arises from these insensible stimuli, which... make us find one direction of movement more comfortable than the other (Leibniz, 1704/1981, p. 53).

Because he believed that perception could be unconscious, Leibniz introduced the term apperception to refer to conscious perception.

At the close of that century, Immanuel Kant (1724-1804) devoted a major section of his last work,Anthropology from a Pragmatic Point of View, to a discussion "Of the ideas which we have without being conscious of them":

To have ideas, and yet not be conscious of them, -- there seems to be a contradiction in that; for how can we know that we have them, if we are not conscious of them? Nevertheless, we may become aware indirectly that we have an idea, although we be not directly cognizant of the same (Kant, 1798/1978, p. 18).

Kant also accepted Leibniz's distinction between unconscious perception and conscious apperception, and further distinguished between the "empirical apperception" of conscious awareness and the "transcendental unity of apperception", or the interconnectedness of all conscious thought.

In the 19th century, Johann Friedrich Herbart (1776-1841), drawing on the views of Leibniz, defined the limen, or sensory threshold, as a mental battleground where various perceptions, themselves mostly unconscious, competed for representation in consciousness:

One of the older ideas can... be completely driven out of consciousness by a new much weaker idea. On the other hand its pressure there is not to be regarded as without effect; rather it works with full power against the ideas which are present in consciousness. It thus causes a particular state of consciousness, though its object is in no sense really imagined (Herbart, 1816/1881, p. 19).

Herbart's concept of thresholds set the agenda for 19th-century psychophysics, and his ideas continue to reverberate in the controversy over subliminal perception (R.F> Bornstein & Pittman, 1992; Dixon, 1971, 1981)). In addition, Herbart's notion that one idea can prevent another from attaining consciousness lies at the foundation of Freud's psychoanalytic theory of repression and defense (Ellenberger, 1970).

Unconscious Inference

Among the most influential advocates of unconscious mental processes was the German physiologist Hermann von Helmholtz (1821-1894). In his Treatise on Physiological Optics, he argued that our conscious perceptions are determined by unconscious inferences concerning the stimulus environment:

The psychic activities that lead us to infer that there in front of us at a certain place there is a certain object of a certain character, are generally not conscious activities, but unconscious ones. In their result they are the equivalent to conclusion, to the extent that the observed action on our senses enables us to form an idea as to the possible cause of this action.... But what seems to differentiate them from a conclusion, in the ordinary sense of that word, is that a conclusion is an act of conscious thought.... Still it may be permissible to speak of the psychic acts of ordinary perception as unconscious conclusions... (Helmholtz, 1866/1968, p. 174; Warren & Warren, 1968).

For Helmholtz, then, perception results from a kind of syllogistic reasoning, in which the major premise consists of knowledge about the world acquired through experience, and the minor premise consists of the information provided by the proximal stimulus. The perception of this proximal stimulus, then, is the conclusion of the syllogism -- except, in the case of perception, we are not aware of the reasoning process. Hence, Helmholtz's phrase, "unconscious conclusions". The fact that the reasoning underlying perception is unconscious is what makes us feel that we are seeing the world the way it really is. But our percepts are not immediate products of stimulation -- they are mediated by unconscious reasoning processes.

The phenomenon of size constancy illustrates Helmholtz's unconscious inferences in operation. Every object in our visual field casts an image of itself on the retinas of our eyes. As a matter of optical physics, the size of the retinal image is proportional to the size of the object, but inversely proportional to the distance between the object and the observer. Thus, objects A and B, which differ in both size and distance from the observer, cast images of the same size on the observer's retina. When an object moves closer to us, the size of its retinal image grows, as from A to A', but we do not perceive the object as growing in size. This is size constancy. In order to maintain size constancy, the visual system takes account of cues to distance, and essentially multiplies retinal size by distance. In Helmholtz' terms, given the knowledge about the size of the distal stimulus as a major premise, and the minor premise that the proximal stimulus cast on the retina is growing in size, perceivers apply the size-distance rule and conclude that the object is coming nearer. We are not aware of making these calculations, of course, and we don't draw the logical conclusion consciously and deliberately. It all just happens automatically, based on knowledge acquired through experience, as part of the proper functioning of the visual system.

Helmholtz' theory of unconscious inference had roots in the earliest accounts of optics and vision, beginning with the Greco-Egyptian astronomer Ptolemy and the Arab mathematician Ibn al-Haytham and running through Descartes, Kant, and Berkeley (for a critical review, see Hatfield, 2002). Nevertheless, both James (1890/1980) and Klein (1977) asserted that Helmholtz later repudiated the notion of unconscious inferences. In fact, his later writings make clear that Helmholtz had reservations only about the term, not the idea.

"Later I avoided that term, 'unconscious conclusions', in order to escape from the entirely confused and unjustified concept -- at any rate so it seems to me -- which Schopenhauer and his disciples designate by this name" (Helmholtz, 1878/1968, p. 220).

Helmholtz reaffirmed the basic idea of unconscious inferences in one of his last essays, republished as part of the second edition (1896) of the Treatise on Physiological Optics:

"...I find even now that this name is admissible within certain limits since these associations of perceptions in the memory actually take place in such a manner, that at the time of their origin one is not aware of it... (Helmholtz, 1894/1968, p. 255).

Despite challenges (e.g., Gibson, 1966; Gibson, 1979), Helmholtz' theory of unconscious inference continues to dominate theoretical discussion of perceptual processes (Hatfield, 2002). Unconscious inference lies at the core of the "constructivist" view of perception, which holds that the pattern of proximal stimulation is not sufficient to yield perception. Rather, perception is an intelligent activity by which the perceiver constructs an internal, mental representation of the world by combining information provided by the stimulus environment with world-knowledge retrieved from memory (e.g., Gregory, 1966, 1970; Hochberg, 1964, 1978; Irvin Rock, 1983, 1995, 1997). Rock, in particular, argued that perception that unconscious (or, at least, unnoticed) cognitive processes occur at four different stages in the perceptual process: form perception, the resolution of stimulus ambiguity, the perception of the relations among the objects of perception, and the construction of perceptual constancies.

Romantic Views of the Unconscious

The pre-Freudian analysis of unconscious mental life reached its apex with Eduard von Hartmann (1842-1906) and his Philosophy of the Unconscious (1868/1931), an extremely popular work whose three volumes, running to more than a thousand pages, went through a total of 12 editions (the last published in 1923). For Hartmann, the universe was ruled by the Unconscious (the initial capital is his) a highly intelligent dynamic force composed of three layers: the Absolute Unconscious, accounting for the mechanics of the physical universe; the Physiological Unconscious, underlying the origin, evolution, development, and mechanisms of life; and the Relative Unconscious, which Hartmann considered to be the origin of conscious mental life. Hartmann's "relative unconscious" is what I prefer to call the psychological unconscious -- a term referring to those mental states and processes which influence our experience, thought, and action outside phenomenal awareness and independent of voluntary control.

Hartmann ascribed a number of specific functions to the Unconscious (Vol. 2, pp. 38-39):

  1. The Unconscious forms and preserves the organism, repairs its inner and outer injuries, appropriately guides its movements, and mediates its employment by the conscious will.
  2. The Unconscious supplies every being in its instinct with what it needs for self-preservation, and for which its conscious thought does not suffice....
  3. The Unconscious preserves the species through sexual and maternal love... and conducts the human race historically steadily to the goal of its greatest possible perfection.
  4. The Unconscious often guides men in their actions by hints and feelings, where they could not help themselves by conscious thought.
  5. The Unconscious furthers the conscious process of thought by its inspirations..., and in mysticism guides mankind to the presentiment of higher, supersensible unities.
  6. It makes men happy through the feeling for the beautiful and artistic production.

We owe to Hartmann the notion, still with us in some quarters today, that the unconscious possesses capacities and powers that are superior to those available to ordinary consciousness:

If we now institute a comparison between the Conscious and Unconscious, it is first of all obvious that there is a sphere which is always reserved to the Unconscious, because it remains forever inaccessible to consciousness. Secondly, we find a sphere which in certain beings only belongs to the Unconscious, but in others is also accessible to consciousness. Both the scale of organisms as well as the course of the world's history may teach us that all progress consists in magnifying and deepening the sphere open to consciousness; that therefore in a certain sense consciousness must be the higher of the two. Furthermore, if in man we consider the sphere belonging both to the Unconscious and also to consciousness, this much is certain, that everything which any consciousness has the power to accomplish can be executed equally well by the Unconscious, and that too always far more strikingly, and therewith far more quickly and more conveniently for the individual, since the conscious performance must be striven for, whereas the Unconscious comes of itself and without effort.... [T]he Unconscious can really outdo all the performances of conscious reason..." (Vol. 2, pp. 39-40).

Hartmann's view of the unconscious is in tune with the romanticism that characterized much European and American literature, art, and music in the 19th century (Whyte, 1960). As a rebellion against both classicism and rationalism, romanticism celebrated nature, sensuality, and feelings -- including what Whyte (p. 132) called "the dark realms of the soul". Jean-Jacques Rousseau (1712-1778) celebrated the unconscious and involuntary role of the will and emotions in behavior; Johann Gottlieb Fichte (1762-1814) asserted that unconscious impulses reflected the intelligence of nature; Friedrich William Joseph von Schelling (1775-1884) asserted that a single unconscious principle organized all of nature; Johann Wolfgang von Goethe (1749-1832), the poet who epitomized the Sturm und Drang school of German literature, celebrated the unconscious roots of human invention and discovery, and asserted that the roots of human nature were in the Unconscious. Arthur Schopenhauer, (1788-1860), the philosopher whose writings caused Helmholtz to regret using the word "unconscious", asserted that our experience of nature was governed by a unconscious and impersonal Will, which is the source of all human pain and suffering, and from which humans must become detached to survive. Hartmann's Philosophy of the Unconscious comes straight out of Schopenhauer, and leads straight into Friedrich Wilhelm Nietzsche (1844-1900), who wrote that consciousness is only the last, and the weakest, stage in organic evolution. And we know what came from Nietzsche: no wonder Helmholtz was worried!

However influential Romanticism was on 19th-century European intellectual life, in the end Hartmann's ideas proved to be too speculative for the first generation of scientific psychologists. Herman Ebbinghaus, discussing Hartmann's book, concluded that "Wherever the structure is touched, it falls apart.... What is true is not new, and what is new is not true" (Shakow, 1930, p. 510)). William James echoed the criticism:

Hartmann fairly boxes the compass of the universe with the principle of unconscious thought. For him there is no nameable thing that does not exemplify it. But his logic is so lax and his failure to consider the most obvious alternative so complete that it wold, on the whole, be a waste of time to look at his arguments in detail (James, 1890/1980, p. 171).

Freud and Psychoanalysis

Despite James's reaction, all of this activity, from Leibniz and Kant to Helmholtz and Hartmann, and beyond, laid the foundation for what Henri Ellenberger , the great historian of psychiatry, called The Discovery of the Unconscious (Ellenberger, 1970). As early as 1868 -- the very year that Hartmann published his Philosophy of the Unconscious -- Henry Maudsley, the father of British psychiatry (after whom the Maudsley Psychiatric Institute in London is named) had noted that "we cannot overestimate the importance of the fact that 'consciousness' is not coextensive with the mind" (Maudsley, 1876/1977, p. xxx). The psychiatric discovery of the unconscious was consolidated with what Ellenberger called a new dynamic psychiatry B the psychiatry of Freud, Jung, and Adler. According to Ellenberger, Leibniz had coined the term "dynamic" to make the point that mental states were active rather than static, and this usage persisted in the hands of Herbart and Fechner. What unites all the dynamic approaches to mental illness is the idea that unconscious mental contents, far from being latent (like unattended stimuli and forgotten memories), actively influence the person's conscious experience, thoughts, and actions. Dynamic psychiatry is based on motives -- not just ideas, but ideas with force behind them; these motives, in turn, are unconscious, so that we do not know why we do what we do.

Based on their clinical observations, for example, Breuer and Freud (1893-1895/1953) concluded that the symptoms of hysteria were produced by unconscious memories of traumatic events (Breuer & Freud, 1893-1895/1953). As they famously put it, "hysterics suffer mainly from reminiscences" (pp. 7).

Our observations have shown... that the memories which have become the determinants of hysterical phenomena persist for a long time with astonishing freshness and with the whole of their affective colouring. We must, however, mention another remarkable fact... that these memories, unlike other memories of their past lives, are not at the patients' disposal. On the contrary,these experiences are completely absent from the patients' memory when they are in a normal psychical state, or are only present in highly summary form. Not until they have been questioned under hypnosis do these memories emerge with the undiminished vividness of a recent event (p. 9).

Freud, of course, went on to elaborate a theory which attributed all conscious experiences, thoughts, and actions to unconscious sexual and aggressive drives, and defense mechanisms (such as repression) unconsciously deployed to control them. When James and Freud met at the famous Clark University conference of 1909, James reportedly told Ernest Jones, a young disciple of Freud and later Freud's biographer, that "the future of psychology belongs to your work".  We do not know exactly what James meant by this remark (Taylor, 1999), or even whether he actually said it. We can suppose that he referred to the analysis of unconscious mental life. In any event, we do know that, for better or for worse, Freud's theories dominated both scientific and popular conceptions of the unconscious for most of the next century.

Consciousness and the Unconscious After Watson

Unfortunately, just when the concept of the psychological unconscious was getting up steam, the behaviorist revolution hit. Interest in consciousness disappeared virtually overnight, and interest in the psychological unconscious went with it. For John B. Watson (1878-1958), the founder of behaviorism, and his comrades in arms, the only way to make psychology truly scientific was to abandon the mental, and focus on behavior. Watson did not deny the existence of consciousness: he only denied that consciousness was subject to scientific investigation. It was bad enough to try to explain behavior in terms of mental states that couldn't be publicly observed; it was even worse to try to explain behavior in terms of mental states that couldn't be privately observed! For Watson, the job of psychology was to discover laws relating publicly observable environmental stimuli with publicly observable muscular and glandular responses (Watson, 1913, 1919). Behavior was construed as a string of innate and conditioned reflexes. Introspection was replaced with verbal reports, though verbal reports, as pieces of behavior, were admitted into the science only when they could be compared with objective stimulus conditions. "The unconscious" was reformulated by Watson merely as the unverbalized (Watson, 1928, pp. 93-115).

Of course, psychological interest in consciousness did not die completely with the triumph of behaviorism. Some psychologists influenced by psychoanalysis maintained an interest in consciousness, as well as the unconscious. Of particular importance here was the movement known as psychoanalytic ego psychology. While Sigmund Freud was primarily interested in the id -- the primitive, irrational, biological drives that he felt were the dynamic core of personality, some of his followers, including his daughter Anna (Freud, 1936/1937), took an interest in the ego, and especially the "conflict-free sphere" of the ego, which enabled the person to form mental representations of external reality (e.g., Hartmann, 1939; Rapaport, 1959). At a time when academic psychology was still dominated by the functional behaviorism of Watson and Skinner, the sensory psychophysics of Stevens, and stimulus-response theories of animal and verbal learning, psychoanalytic ego-psychologists were publishing investigations of cognitive development in infants and children, memory for connected discourse, attributions of causality, individual differences in cognitive style, mental imagery, and preconscious processing. Much like the monks of the Middle Ages, the ego psychologists held fast against the behaviorist onslaught, preserving what was most interesting in psychology until psychologists were ready to study the mind again.

There were others besides psychologists who prefigured the cognitive revolution and the resurgence of interest in consciousness. The entire Gestalt movement in perception was fundamentally concerned with the nature of conscious experience. (I. Rock & Palmer, 1990). McDougall's hormic psychology argued against Watson that behavior has a purpose, and is not just a response to stimulation; for McDougall, each episode of mental life began with a thought, continued with an intention, and ended with a feeling (McDougall, 1923). Woodworth's (1938) work on the span of apprehension is essentially an investigation of the limits of consciousness, and presaged work on attention that heralded the cognitive revolution (Woodworth, 1938). Even earlier, Wallas's analysis of thinking had contrasted an unconscious stage of incubation with the conscious stages of preparation,illumination, and verification (Wallas, 1921). Still, a full-scale revival of academic interest in consciousness had to wait until behaviorism was overthrown by the cognitive revolution. And when the cognitive revolution happened, consciousness was there in the thick of it

In the 1950s and 1960s, research on selective attention (Broadbent, 1958), primary or short-term memory (Atkinson & Shiffrin, 1968; Waugh & Norman, 1965), and especially on mental imagery (Richardson, 1969; Segal, 1971; Sheehan, 1972) legitimized the study of consciousness. Somewhat ironically, however, the early cognitive psychologists hardly ever used the term consciousness itself. This is a result, I think, of four tendencies described by the philosopher Owen Flanagan (1992).

First, as a methodological holdover from behaviorism, is a positivistic reserve reflecting our persisting reluctance to use mentalistic language. Behavior is our window on the mind, but many psychologists remain more comfortable talking about behavior than about mental states.

Second, cognitive psychology prefers a piecemeal approach that assumes that big problems like consciousness can be solved by working up from the bottom on smaller problems like attention and short-term memory. At a cocktail party, one of the world's leading cognitive psychologists once told me proudly that he had written a dozen books without ever using the word consciousness. A focus on discrete phenomena, without addressing the big picture, essentially marginalizes the subject of consciousness. Compare, in undergraduate psychology curricula, the number of courses on consciousness to the number of courses on perception, memory, intelligence and problem-solving, or judgment and decision-making.

The study of consciousness can survive piecemeal approach by investigators suffering hangovers of positivistic reserve, but two more trends described by Flanagan are more dangerous.

One is the epiphenomenalist suspicion: the idea that consciousness is the endproduct of cognitive functioning, but plays no causal role in human experience, thought, and action. From this point of view, we are conscious automata -- zombies who know what we're doing, perhaps, but zombies nonetheless.

From this stance it is only a short step to conscious inessentialism -- the view that understanding consciousness is simply not important to understanding how the human mind works. Within cognitive psychology (and cognitive science generally), proponents of "computational functionalism" seek to describe human cognition in terms of the functional relations between stimuli and responses. This may sound like a throwback to Watsonian and Skinnerian behaviorism, and in some respects it is -- the principal difference being that the computational functionalists are interested in the information-processing structures that mediated between stimulus and response, while Watson and Skinner were not. But computers process information, and hardly anyone raises the question of whether they are conscious. Accordingly, it may seem to some that we might have a complete description of human information processing without ever getting to the topic of consciousness at all.

If the cognitive revolution laid the basis for the revival of interest in conscious mental life, the attitudes just described sowed the seeds for a revival of interest in unconscious mental life as well. There were other sources as well. At the very beginning of the cognitive revolution, the linguist Noam Chomsky (1957) argued that human language was mediated by "deep" grammatical structures which are inaccessible to conscious introspection, and can be known only by inference (see, especially, Chomsky, 1980)). Along the same lines, the philosopher Jerry Fodor argued that many mental functions, such as visual perception, were mediated by dedicated structures which were impenetrable by conscious awareness and control (Fodor, 1983). As noted earlier, cognitive approaches to perception, as exemplified by the work of Irvin Rock on visual illusions, invoked a version of Helmholtz's notion of unconscious inference to explain both how we see the world the way we do (Irvin Rock, 1983, 1997). Finally, the classic multistore model of memory, which made a place for consciousness in attention and primary memory, made a place for the psychological unconscious in its list of preattentive, or preconscious, information processing functions (Atkinson & Shiffrin, 1968; Waugh & Norman, 1965).

Four Roads to the Unconscious

In the final analysis, however, the revival of interest in the psychological unconscious was not just a byproduct of psychologists' squeamishness about consciousness and mentalistic talk, and it was more than an expansion of footnotes in theories of perception or thinking. Interest in the psychological unconscious now runs wide and deep within psychology. This happy state of affairs is the end product of at least four quite independent strands of investigation, which together converge on our modern conception of unconscious mental life:

  • The distinction between automatic and controlled mental processes.
  • Cognitive neuropsychology, especially the study of implicit memory and blindsight in brain-damaged patients.
  • A revival of interest in "subliminal" perception.
  • Last, but not least, increased interest in the phenomena of hypnosis, and in particular priming effects observed in posthypnotic amnesia, hypnotic blindness, and the like.

Automaticity

Now we've already seen, back in the lecture on short-term and working memory and attention, one aspect of unconscious mental life, which has to do with the distinction between automatic and controlled processes.  You'll remember that automatic processes are inevitably evoked by the appearance of some stimulus in the environment.  Once evoked, they run incorrigibly until completion: they cannot be stopped.  They are executed efficiently, in the sense that they consume little or no cognitive resources.  And they can be processed in parallel, which means that they don't interfere with other ongoing mental activities.

Controlled processing is often identified with conscious processing.  We're aware of what we're doing, and we can control it voluntarily.  Automatic processing, by contrast, seems to be unconscious in the strict sense of the term.  Automatic processes operate outside of our phenomenal awareness.  And they also operate outside of our voluntary control.

One research tradition contributing to the modern interest in the psychological unconscious is based on a distinction between "automatic" and "strategic" mental processes. Skilled reading provides one example of automaticity: we recognize certain patterns of marks on the printed page as letters, and certain patterns of letters as words, and decode the meanings of words in light of the words around them, but we rarely have any conscious awareness of the rules by which we do so. It just happens, as an automatic consequence of having learned to read. The power of these processes is illustrated by the color-word effect discovered by Stroop (MacLeod, 1991; see also MacLeod, 1992; 1935). In the basic "Stroop" experiment, subjects are presented with a list of color names printed in different colors, and are asked to name the color in which each word is printed. This task is easy if the ink-color matches the color name (e.g., the word yellow printed in yellow ink); but if the word and its color do not match (e.g., yellow printed in green ink), it is very hard. Despite the subjects' conscious intention to name ink colors, and to ignore the words themselves, they cannot help reading the color names, and this interferes with naming of the colors. It just happens automatically.

According to traditional formulations (e.g., Schneider & Shiffrin, 1977), automatic processes are inevitably engaged by the appearance of specific environmental stimuli, independent of the person's conscious intentions. Once invoked, they proceed incorrigibly to their conclusion. In theory, their execution consumes no attentional resources, and for this reason (and perhaps also because they are executed very rapidly) they leave no traces of themselves available to conscious recollection. Hasher and Zacks (1979) have further suggested that automatic processes do not improve with training or feedback. Furthermore, assuming that the person is neurologically intact, once they have been acquired automatic processes are invariant across age, education, socioeconomic status, race, and ethnicity.

The unconscious inferences discussed by Helmholtz (1866/1968) probably occur automatically. Certainly we are not aware of making these inferences, and we do not intend to make them -- they just seem to happen in the course of perceiving, and we are consciously aware only of their endproduct. In theory, some mental processes are innately automatic, while other processes become automatized only after extensive practice with a task. More recently, some hypothesized properties of automaticity have been called into question by revisionist, memory-based theorists (e.g., Logan, 1997). But even these revisionist views agree that some mental processes are unconscious in the strict sense of the term: they are inaccessible to phenomenal awareness under any circumstances, and can be known only by inference from task performance.

The concept of automaticity, originating in the rather narrow confines of research on visual search and pattern recognition, has come to play an increasingly powerful role in personality, social, and clinical psychology. The general argument is that some attitudes, impressions, and other social judgments, as well as aggression, compliance, prejudice, and other social behaviors, are mediated by automatic processes which operate outside phenomenal awareness and voluntary control (e.g., Bargh & Chartrand, 1999; Kirsch & Lynn, 1999a; Wegner & Wheatley, 1999).  To some extent, which I have come to think of as the automaticity juggernaut seems to represent a reaction to a cognitive view of social interaction which seems, to some, to inappropriately emphasize conscious, rational, cognitive processes, at the expense of the unconscious, irrational, emotive, and conative. In addition, the popularity of automaticity seems to represent a reversion to earlier, pre-cognitive, situationist views within social psychology.  After all, the concept of automaticity is at least tacitly modeled on innate stimulus-response connections such as reflexes, taxes, and instincts, as well as those acquired through classical and instrumental conditioning. The automaticity juggernaut is not exactly a reversion to Skinnerian behaviorism, because it entails internal mental representations and processes intervening between stimulus and response, but it is close: if the cognitive processes underlying social cognition and social behavior are indeed largely automatic, then not too much thought has gone into them.

My own view is that automatic processes play an important role in mental life, they are by no means decisive for thought and action.  In particular, the fact that some mental processes occur automatically does not undermine our usual concepts of free will.

In a sense, automatic processes are exemplified by the unconscious inferences that Helmholtz thought lay at the root of perception.  So, for example, the relative distance of two objects from the observer can be inferred from the relative differences in the size of the images that the objects cast on the retina.  We make this inference by employing the size-distance rule, that distance is an inverse function of the size of the retinal image, but we're not conscious of making this inference.  The inference is made automatically, by our visual system.  More broadly the idea is that automatic unconscious processes are responsible for generating conscious mental contents -- our conscious sensations, percepts, and memories, our conscious thoughts, feelings, and desires are to a large extent generated by these unconscious automatic processes.  Both traditional and revisionist approaches to automaticity assume, at least tacitly, that the mental contents upon which these processes operate are accessible to conscious awareness. So automatic processes can be evoked by conscious thoughts and feelings, and they can create conscious thoughts and feelings.  The next question is: Can thoughts and feelings themselves also be unconscious?

The Cognitive Unconscious

Apparently, the answer is yes.  It is now clear that our experiences, thoughts, and actions can be influenced by mental contents -- percepts, memories, thoughts, feelings, and desires, of which we are unaware. Compelling evidence for this proposition began to accumulate in the 1960s, as cognitive psychology turned into cognitive neuropsychology, and researchers began to see evidence of the psychological unconscious in the behavior of brain-damaged patients.

Less than 100 years after Immanuel Kant raised the question of whether unconscious ideas were possible, Peirce and Jastrow showed that they in fact existed, in their experiment on subliminal perception.  You'll remember that in this experiment, subjects were asked to identify which of two stimuli were the brighter, or the heavier, and then the difference between them was systematically reduced until the subjects could not consciously detect a difference any longer.  Yet when forced to choose which stimulus was the brighter or the heavier, they were right significantly more often then we'd expect by chance.  They were not consciously aware of the difference between the two stimuli, but that difference registered in their sensory system somehow and affected their choices, their guesses, outside of conscious awareness.

Implicit Memory

In modern psychology, initial evidence for the existence of unconscious mental contents, as opposed to unconscious mental processes, was provided by laboratory studies of patients with the amnesic syndrome (sometimes known as "Korsakoff's syndrome" after the neurologist who first described it in 1889) caused by bilateral damage to the hippocampus and other structures in the medial portion of the temporal lobes or, alternatively, to the mammillary bodies and related structures of the diencephalon, which includes the thalamus and hypothalamus. 

  • On clinical observation, such patients show a dense anterograde amnesia for "postmorbid" events that occurred after the brain damage: after only a few moments of distraction, they cannot consciously remember events that have occurred just recently.
  • They may also show some degree of retrograde amnesia, covering "premorbid" events that occurred prior to the brain damage, but this point is controversial and need not concern us here.

As the study of these patients shifted from clinical description to controlled laboratory investigation, however, it became apparent that the events apparently covered by the amnesia nonetheless influenced the patients' ongoing experience, thought, and action.

We discussed earlier the pioneering study by two British psychologists, Elizabeth Warrington and Lawrence Weiskrantz, in which amnesic patients studied a list of words and then received tests of free recall and recognition.  Compared to the control subjects, the amnesic patients were much less likely to produce items from the studied lists on those tests.  Compared to non-amnesic controls, the amnesic patients showed profound deficits on standard tests of recall or recognition: this is the basic phenomenon of anterograde amnesia. And that's pretty much the definition of amnesia -- an inability to consciously remember events from the past.  However, when presented with word fragments or word stems, and asked to guess the corresponding word, amnesic patients were much more likely to respond with list items.  In fact, their performance on the completion test was no different from that of controls.  This is known as a priming effect.  Apparently studying the word list left a trace in memory.  Both amnesics and controls performed extremely well on this "partial information" test -- a phenomenon known as repetition priming because the cue provided at the time of testing is a repetition (at least in part) of an item presented during the study phase. The amnesic patients didn't have conscious access to that memory trace, which is why they performed so poorly on tests of free recall and recognition.  But traces of the studied words had been encoded anyway, and remained in storage, and were able to influence the amnesic patients' performance on the stem- and fragment-completion tests.  Priming shows that some memory of the items encountered in the study phase was retained, despite the patients' failure of conscious recall and recognition.

Priming is clearly an effect of memory: if some trace of studied items did not remain in storage, there would be way for the subject to guess the word represented by stem at better than chance levels. Interestingly, the amnesic patients had little or no conscious recollection of the study phase, and treated the partial information test as a kind of guessing game. "Guessing" proved to be an important element in their performance. Later research showed that when amnesic patients were specifically instructed to complete word stems and fragments with items from the studied list, they did not perform as well as controls (e.g., Graf, Squire, & Mandler, 1984). Nevertheless, the fact that they displayed levels of repetition priming equivalent to those of controls -- indeed, the fact that they displayed significant levels of priming at all -- showed that, in some sense, they still retained some memories of their previous experience.

Based on results such as these, Daniel Schacter and his colleagues (1987) have distinguished between two expressions of memory: explicit and implicit.

  • Explicit memory refers to the conscious recollection of a past event, as indicated by performance on recall or recognition tasks.  
  • Implicit memory refers to the effect of a past event on some task which does not require conscious recollection, like priming effects.  Implicit memory refers to any effect of a past event on the individual's experience, thought, or action. 

Repetition priming effects, in which prior exposure to a word like table makes it more for a person to complete the stem tab__ with table than with tablet or taboo, or prior exposure to a word like assassin makes it easier for a person to complete a fragment like __ss___ss__ at all, are good examples of implicit memory. They obviously depend on memory, but the task does not refer to the past, nor, logically, does it require conscious recollection of any past event. All the subject has to do is to generate an acceptable word that fits in the spaces provided. The sparing of implicit memory in amnesia shows that some representation of a prior event has been encoded and stored in memory, and influences ongoing experience, thought and action, even though that event cannot be consciously remembered. Implicit memories are unconscious memories.

Subsequent research has confirmed the essential findings of Warrington and Weiskrantz -- that explicit conscious, and implicit unconscious memory can be dissociated in the amnesic syndrome.  That is to say, implicit memory occurs in the absence of explicit memory.  But this dissociation between explicit and implicit memory has also been observed in a wide variety of other conditions.

  • For example, the amnesia that occurs as an incidental byproduct of electroconvulsive therapy in the treatment of depression.  
  • Implicit memory is also preserved for information presented during surgery under general anesthesia,
  • and also conscious sedation, another procedure often used in outpatient office procedures. 
  • Some dementing illnesses, like Alzheimer’s disease, profoundly affect explicit memory but leave implicit memory intact, at least in the early stages of the illness. 
  •  And mental patients with various forms of dissociative disorder, like so-called psychogenic amnesia, fugue, or multiple personality disorder, will often show priming affects created by past events which they can't consciously recollect.  
  • In posthypnotic amnesia, hypnotized subjects come out of hypnosis and can't remember the things they did or the things they experienced while they were hypnotized; yet they'll show priming effects for a list of words that they studied while they were in that state.

Dissociations between explicit and implicit memory can also be observed in “normal” memory. 

  • For example, memory problems are a feature of normal aging, but they affect explicit memory much more than implicit memory.
  • We don’t know much about implicit memory in young children, but the evidence suggests that they display implicit memory before they have the capacity for conscious recollection. 
  • Finally, explicit and implicit memory can be dissociated even in an ordinary verbal-learning experiment.  For example, explicit memory is profoundly affected by the level of processing received by an item at the time of encoding (what I call the Elaboration Principle). But implicit memory is much less affected by this manipulation.

In all these cases, implicit memory is unconscious memory.  It's a mental representation of past experience, stored in memory, which influences the person's ongoing experience, thought or action, in the absence of conscious recollection.

The concept of implicit memory has received a huge amount of attention in the field. A whole industry has developed around implicit memory, involving amnesic and demented neurological patients; dissociative disorders such as psychogenic amnesia, fugue, and multiple personality; children and the healthy aged; depressed patients receiving electroconvulsive therapy; surgical patients receiving general anesthesia or conscious sedation; and even college students who have all their wits about them (e.g., Roediger & McDermott, 1993; Schacter, Chiu, & Ochsner, 1993). Rather than focusing on existence proofs -- which is still somewhat the situation for subliminal perception (see, e.g., Draine & Greenwald, 1998) -- the implicit memory literature assumes at the outset that the notion of unconscious, dynamically active memories is fundamentally valid, and focuses instead on the mechanisms by which unconscious memories are dissociated from conscious recollection.

Still, there are some important issues that have to be addressed by further research. For example, almost all the evidence on implicit memory has been collected within a single narrow paradigm, repetition priming, leading to theories of implicit memory which emphasize relatively low-level perceptual processes. But semantic priming occurs too, not just in posthypnotic amnesia but in organic amnesia as well, suggesting that these perception-based theories are not adequate to the phenomenon. Similarly, while explicit and implicit memory are dissociable, they also interact, requiring revisions in theories which hold that these two expressions of memory are mediated by separate memory systems in the brain (Kihlstrom, 1998c).

Despite these and other persisting questions, the general acceptance of the distinction between explicit, conscious and implicit, unconscious expressions of memory opens the door to extensions of the explicit-implicit distinction to other domains of mental life, across the whole of Kant's trilogy of mind - -cognition, emotion, and motivation.

Implicit Perception

Just as implicit memory refers to the influence of past events that cannot be consciously remembered, so implicit perception refers to the influence of events in the current stimulus environment that cannot be consciously perceived.  Explicit perception is conscious perception, as exemplified by our ability to consciously detect a stimulus in the environment; to perceive its distance, motion, or form; or to identify the object and to categorize it.  Implicit perception, by analogy with implicit memory, refers to any effect of a current event -- an event in the current stimulus environment -- on the individual's experience, thought, or action, in the absence of conscious perception.

Perception without awareness can be observed in neurologically intact subjects, in the form of "subliminal" perception. Herbart's proposal that there is a limen, or threshold, which must be crossed before a stimulus could be consciously perceptible stimulated the development of psychophysicists. However, Leibniz's notion of petites perceptions implied that there were percepts below the limen. If these percepts actively influenced behavior, as Leibniz argued, then perhaps there was no limen at all. The quest to document subliminal perception began in the 1880s with the Peirce & Jastrow (1884) experiment, resulting in the accumulation of a substantial body of positive evidence (Adams, 1957; Dixon, 1971).

Unfortunately, these experiments were subject to vigorous and often persuasive criticism, mostly on methodological grounds (Eriksen, 1960; Goldiamond, 1958). Chief among these criticisms was the argument that the investigators had improperly determined the thresholds in question, so that what appeared to be unconscious perception might well have been conscious perception after all. It did not help, either, that before World War II, the question of subliminal perception was raised mostly under the influence of psychoanalysis. After the war, the idea was revived as part of Bruner's "New Look" in perception (Bruner & Klein, 1960), which was also influenced, to some degree, by Freudian psychoanalysis. There was also a vigorous debate, stimulated by Vance Packard's (1957) book,The Hidden Persuaders, over the role of subliminal messages in advertising. The New Look was a forerunner of the cognitive revolution, but the experimental psychology of the 1950s and early 1960s, dominated as it was by positivism and behaviorism, was not much interested in cognition -- and it was certainly not interested in anything tainted by psychoanalysis. Accordingly, the substantial body of evidence reviewed by reviewed by Adams and Dixon was largely dismissed and ignored.

That might have been the last we heard of subliminal perception, except that in the early 1980s Anthony Marcel (1983a; Marcel, 1983b) presented solid evidence of the phenomenon in research on semantic priming effects on lexical decision. In the lexical decision paradigm, the subject must decide whether a string of letters is an actual word in English. This is normally an easy task, but it is even easier if the target word is preceded by another, semantically related, word (Meyer & Schvaneveldt, 1971). This phenomenon is known as semantic priming, as distinct from the repetition priming studied in amnesic patients by Warrington and Weiskrantz (1970). Marcel's innovation was to present the prime surrounded by "masking" stimuli, so that it was rendered invisible to the subjects (the targets, though, remained completely visible). In masking, stimuli are presented at "supraliminal" intensities that would be sufficient to allow them to be perceived under ordinary circumstances.  This is different from the Peirce and Jastrow case, where the stimuli were presented at intensities that were too low to be consciously perceived.  So the stimuli themselves are nominally supraliminal, above the threshold for conscious perception.  But they're followed by a masking stimulus that effectively renders them invisible.  This masking stimulus is typically a visual pattern like a string of hashmarks. 

In his experiments, Marcel carried out a series of studies of masked semantic priming in which words like doctor facilitated lexical decisions of semantically related targets like nurse.  All the subjects had to do was to determine whether a letter string like “n-u-r-s-e” was a legal English word.  Now ordinarily, you'd think that was no big deal, because “doctor” and “nurse” are semantically related, and so presentation of one would be expected to prime processing of the other.  But because of the masking stimulus, the subjects were not consciously aware of the prime.  Attesting to the effectiveness of the masking procedure, detection of the primes was at chance levels, even when they had been presented 20 times. Yet the word was perceived unconsciously.  And we know it was perceived unconsciously because masked presentation of the word “doctor” primed the processing of the semantically related word “nurse”.  This was the case even after a single presentation of the prime, and this effect grew with increasing numbers of repetitions. So implicit perception goes beyond merely subliminal perception, and it goes beyond the simple question of whether a stimulus is detectable.  In the Marcel experiment, the subjects were actually reading words and understanding them, analyzing them for meaning, despite the fact that they weren't consciously aware of the words.  Masked semantic priming gives very good evidence for implicit perception – for unconscious perception.

A major strength of Marcel's research lay in his careful attention to the determination of thresholds. His report raised a firestorm anyway, with a number of critics essentially repeating the criticisms that had been raised against the earlier generations of research (e.g., Holender, 1986). I think this response occurred mostly because of the historical association of subliminal perception with psychoanalytic theory. But another reason was the simple fact that the cognitive theories of the time tended to describe cognition in terms of a series of ever more complicated processes, and had no room in them for the possibility that the meanings of words could be analyzed unless conscious attention was paid to them. The criticisms had the character of the apocryphal entomologist who found a bug he couldn't classify, so he stepped on it. Despite these criticisms, Marcel's findings were soon confirmed by others (e.g., Balota, 1983; Fowler, Wolford, Slade, & Tassinary, 1981), some of whom addressed the matter of thresholds even more assiduously then he did (e.g., Greenwald, Klinger, & Liu, 1989). These days, a wealth of evidence supports the validity of subliminal perception, defined broadly as the influence of stimuli that are too degraded by their conditions of presentation to be accessible to conscious perception. The debate over subliminal perception is not so much over whether it occurs, as over its extent.

As was the case with memory, it's now understood that explicit and implicit perception, conscious and unconscious perception, can be dissociated in a wide variety of conditions.  We've already talked about so-called subliminal perception, and also masked priming -- as in the Marcel experiment.  

The priming that occurs for stimuli presented during general anesthesia could easily be thought of as an example of implicit perception, because -- of course -- an adequately anesthetized patient isn't aware of anything that's going on at the time it's happening.

Implicit perception has also shown to be preserved in a number of neurological syndromes, where patients don't have the conscious experience of seeing or perceiving but are still able to respond appropriately to a perceptual stimulus. 

  • Blindsight patients have damage to area V1, the primary visual cortex in the occipital lobe (Weiskrantz, 1986)Such patients experience a scotoma -- a portion of the visual field where they have no visual experience. When a visual stimulus is presented to that portion of the visual field corresponding to their scotoma, they see nothing at all. Yet their conjectures about the visual properties of the stimulus can be amazingly accurate. For example, the patient known as R.B., studied by Weiskrantz, made guesses about the presence, location, form, movement, velocity, orientation, and size of objects that were more accurate than would be expected by chance alone. The latest evidence suggests that blindsight patients can even make rather difficult discriminations among facial expressions of emotions, even though they cannot consciously see the faces in question (DeGelder, Vroomen, Pourtois, & Weiskrantz, 1999). Just as amnesic patients are influenced by past events that they cannot consciously remember, so blindsight patients can respond to visual stimuli in their current environments that they cannot consciously perceive.
  • In visual neglect, patients with another form of brain damage, affecting the temporo-parietal portion of one hemisphere (usually the right), but sparing the primary sensory and motor cortices, appear not to see -- they appear not to be attentive to stimuli that are present in one half of their visual field -- typically the left half.  Yet careful testing shows that they're responsive to features of those stimuli, even though they're not paying any attention to them and don't see them.
  • Patients with prosopagnosia can't recognize the faces of people who ought to be familiar to them, whether it's their spouses, or their children, or the President.  Yet if you measure their physiological responses to these faces, you'll see that they respond physiologically in a different manner to objectively familiar faces, than they do to the faces of people who are completely unknown to them.
  • Conscious perception is impaired in a group of psychiatric disorders known as the conversion disorders, which used to be known as “hysteria”: patients claim to be unable to see or hear, or feel tactile sensations, despite the fact that there is nothing wrong physically with their sensory systems.  Careful testing in such patients often shows that implicit perception is unimpaired. 
  • In a somewhat similar manner, hypnotized subjects can receive suggestions for blindness, for deafness or for tactile anesthesia, after which they don't see, or hear, or feel things that they're supposed to see, or hear, or feel.  Yet, we can show through careful testing, using priming paradigms and other kinds of procedures, that they're actually responsive to these stimuli outside of conscious awareness. 

Implicit perception is also preserved when stimuli are presented to normal subjects, subjects who are neurologically intact and don't have any kind of psychiatric disorder, but the stimuli are presented outside the field of attention. 

  • In “parafoveal” vision and dichotic listening, we can get priming effects even though the subjects are not consciously attending to stimuli presented outside the field of their attention.
  • You can also get these effects even when subjects are paying attention to the very field in which the stimulus is presented – they’re just not paying attention to that particular stimulus.  These phenomena are collectively known as “cognitive blindness”. 
    • "Inattentional blindness” is exemplified by a “selective looking” experiment in which subjects who were paying attention to a ball game failed to notice a man in a gorilla suit walking across the court. 
    • Similar lapses of attention occur under conditions of “rapid serial visual presentation”.  Even so, in at least some of these instances subjects will show priming effects indicating that the unnoticed events have been processed outside the scope of conscious attention.
      • In repetition blindness, subjects fail to notice that a stimulus had been repeated.  
      • In the attentional blink, they fail to notice a stimulus presented immediately after one that they’ve detected. 
      • In change blindness, they don’t notice that a stimulus has been altered. 

So in a wide variety of conditions people are found to be responsive -- they show priming effects and other kinds of responses -- to stimuli of which they're not consciously aware.  That's the essence of implicit or unconscious perception.

The distinction between implicit perception and implicit memory is not always easy to make, because both phenomena are revealed by post-exposure priming effects of various sorts -- that is, by performance on nominal tests of implicit memory (J.F. Kihlstrom, 1996a). Schacter's (1987) seminal review of the implicit memory literature referred to both Weiskrantz's (1986) studies of blindsight and Marcel's (Marcel, 1983a, 1983b) experiments on subliminal perception. In my view, however, the term implicit memory should be reserved for cases where there is no question that the stimulus event was consciously perceived at the time of encoding. Where there is no conscious awareness at the time of encoding, then we can consider priming effects as evidence of implicit perception: Prior perception is logically implied by spared implicit memory. Thus, on the assumption that adequately anesthetized patients are really unaware, during their procedures, of what our medical colleagues disarmingly call "surgical stimuli", evidence collected in the recovery room of priming effects attributable to events presented during surgery constitutes evidence of implicit perception, not just implicit memory. By contrast, priming effects observed in conscious sedation, where subjects are fully aware of the study trials, is a pure implicit memory effect.

Adopting the implicit-explicit distinction may help resolve a persisting controversy over the scope of unconscious perception. For example, Greenwald (1992) and his colleagues have asserted that subliminal perception is analytically limited B some semantic processing of a subliminal stimulus is possible, but not too much of it. On the other hand, other enthusiasts of subliminal stimulation, including those who employ subliminal techniques in advertising and psychotherapy, assume more processing than Greenwald's arguments would allow. For example, the psychoanalyst Lloyd Silverman (Silverman, 1976; Silverman & Weinberg, 1985) reported that subliminal symbiotic stimulation B by which he meant subliminal presentation of the phrase Mommy and I are one -- had profound effects on the behavior of both normal subjects and psychiatric patients. For this to be true, subliminal perception would have to possess almost Hartmanian cognitive scope, extending to complex semantic analyses far beyond the relatively simple ones permitted under Greenwald's view.

My own view is that Greenwald is right, so far as truly subliminal stimulation goes (J.F. Kihlstrom, Barnhardt, & Tataryn, 1992b), and arguments to the contrary are based more on the Romantic or psychoanalytic notions of the unconscious that I discussed earlier. On the other hand, there are lots of ways to render percepts implicit, and how it is done may make a big difference to what you can do with them. Merikle and Reingold (1988) have usefully distinguished between the objective threshold, where all response to the stimulus drops to chance levels, and the subjective threshold, where the subject simply does not experience the stimulus consciously. Experiments such as Greenwald's experiments take subjects as close to the objective threshold as they can get (Draine & Greenwald, 1998; Greenwald, Draine, & Abrams, 1996; Greenwald et al., 1989), and with such degraded presentations it is not surprising that perception, and memory encoding, are analytically limited. When presentation conditions get closer to the subjective threshold, or with the parafoveal presentation of supraliminal stimuli, more extensive analyses, still outside awareness, might be possible. In hysteria and hypnosis, where the stimulus is in no sense degraded and lies within the scope of focal attention, the possibilities for unconscious analysis might be virtually unlimited.

Implicit Learning

Apparently, subjects can also learn unconsciously, in the sense that new semantic and procedural knowledge acquired through experience can affect their ongoing behavior, without their being aware of what they've learned or even the fact that they've learned anything at all.  Continuing the elaboration of the explicit-implicit distinction to other domains, we can define explicit learning as the person's conscious access to knowledge that he or she has acquired through experience.  Whether that knowledge is semantic knowledge, like the meanings of words or some historical fact, or procedural knowledge, some skill or rule that's been learned.  Implicit learning, then, refers to any effect of this new knowledge, acquired through experience, on the person's experience, thought, or action, even though the person has no conscious awareness of that knowledge. 

As it happens, the term "implicit learning" antedates implicit memory, having been coined by Arthur Reber in 1967. In his now-classic experiments, Reber (1967) asked subjects to memorize a set of letter strings, such as

MSSSSV, MVRXVS, and VXVRXV.

Reber then informed his subjects that the strings they had studied had all been by generated by a particular set of rules -- namely a "Markov process" artificial grammar.  When asked to describe the grammar, none of Reber's subjects could do so. Nevertheless, the subjects were able to distinguish between new letter strings that conformed to the grammar, such as

MSVRXR and VXVRXV,

and other strings that did not, such as

MSRVRX and VXMRXV.

Their performance on the discrimination task revealed that the subjects had acquired knowledge of the grammar, without being able to articulate the grammar in question. They had obviously learned the grammar, at least to some extent, through their exposure to the letter strings.  And they were able to use this knowledge to distinguish between grammatical strings and ungrammatical strings, but they were unable to specify what the grammatical rules were.  They acquired new knowledge, knowledge of the grammar; and that knowledge affected their behavior; but they had no conscious access to the knowledge they had acquired through experience.  

Implicit learning of this sort has also been observed in a wide variety of other paradigms, including classical and instrumental conditioning, control of complex systems, and the learning of categories and sequential relationships (for a review, see Berry & Dienes, 1993). In each of these cases, people's behavior is altered by prior experience -- the classical definition of learning -- even though they are unaware of what they have learned.

According to Reber, the learning displayed by his subjects was implicit in two senses:

  1. they acquired new knowledge incidentally, without intending to do so (they had been asked to memorize letter strings, not to induce a set of rules);
  2. more important, they did not know what they knew (they lacked the ability to consciously articulate the grammar in question).

In his view, the learning of artificial grammars is a laboratory model of the process by which natural language is unconsciously acquired and used. We speak and understand our native language without consciously reflecting on the grammatical rules that our language follows, and we learn these rules simply by virtue of exposure to our native language. Similarly, linguists operating in the Chomskian tradition hold that syntax language acquisition occurs unconsciously and automatically by virtue of a language-specific cognitive module that evolved specifically for that purpose. In Reber's view, however, implicit learning is mediated by a general-purpose learning module that is not specific to language, and which allows the organism to pick up a wide variety of regularities in the environment. Language acquisition, far from being unique, is a special case of this general learning process.

As with implicit perception, the border between implicit learning and implicit memory is a little vague (Kihlstrom, 1996a). Of course, this is as it should be: memory provides the cognitive basis for learning in the first place, and whatever is learned has to be stored in memory. Both implicit memory and implicit learning meet the definition of "any change in experience, thought, and action attributable to a past event, independent of conscious recollection of that event". But the problem goes deeper than that. In the first place, priming effects generally result from a single exposure to the priming event -- a more or less discrete episode in memory. However, implicit learning generally occurs over a sequence of trials -- each constituting a different episode in memory.

Of course, the learning episode is also an experience in the subject's life. When normal subjects learn an artificial grammar, they certainly remember being asked to study the sample strings, and they may even remember the strings themselves, even if they are unaware that the strings follow some rule, or that they have unintentionally abstracted some rule from the studied items. By contrast, when brain-damaged amnesic patients acquire new patterns of behavior from experience, they are amnesic for the whole learning experience. For example, the famous patient H.M. was able to acquire a number of motor skills, as shown by progressive improvements in such tasks as pursuit-rotor learning and mirror-tracing; but despite these improvements, he failed to recognize these tasks as ones he had performed previously (Milner, Corkin, & Teuber, 1968). In amnesia, the occurrence of implicit learning also gives evidence of preserved implicit memory, but I prefer to reserve the term implicit memory for effects that occur in the absence of conscious memory for the original experience. By the same token, I reserve implicit learning for new declarative and procedural knowledge are acquired through learning experiences, in the absence of conscious awareness of what has been learned.


Reber's experiments showed that subjects could unconsciously learn a fairly complex, artificial grammar.  Now, they were certainly aware of memorizing the letter strings at the time they were presented to them for study; so this isn't a case of unconscious perception, and they were obviously able to remember the letter strings themselves; so this isn't a case of unconscious memory, either.  It's a case of unconscious learning.  They've acquired knowledge of the grammar, through experience -- through exposure to grammatical letter strings -- but they're not aware of what they've learned.

Now, you can think of Reber's artificial grammar test as a kind of categorization task.  Subjects are asked to classify strings as grammatical or ungrammatical, and similar findings have been made in other domains of categorization.  Subjects are able to classify two objects as similar, or as belonging to the same category, even though they are not able to say what that category is, or what features are critical for category membership.  Subjects are also able to detect the co-variation (or correlation) between two events, picking up on the association between them, much in the manner of classical conditioning.  They're consciously aware of the two events, but what they're not consciously aware of is the fact that the two events are associated.  Yet they behave as if they are.  They've learned the association between the events, even though they're not consciously aware of that association.  Similarly, subjects can pick up information about the sequencing of events -- even though, when you ask them, they don't think the events are occurring in any particular sequence.  And finally, subjects can learn to control complex systems, even though they can't tell you the rule that relates inputs to the system to outputs from the system.  In each of these cases, subjects have acquired new knowledge through experience -- that's the definition of learning; but they're unaware of the knowledge they have acquired.  That's why the learning is implicit, or unconscious, not explicit and conscious.

At this point, a small industry has developed around implicit learning -- culminating in the publication of a whole "handbook" devoted to the topic (Stadler & Frensch, 1998). Still, it has to be said that the claim of implicit learning remains controversial even after more than 30 years of work. There is a continuing debate over whether implicit learning is really unconscious in any meaningful sense of the term. It just may be too much to expect subjects to be able to articulate an entire Markov process, but subjects might be consciously aware of just enough of the system to permit them to discriminate at above-chance levels between grammatical and ungrammatical strings.

There is also a question about the format in which the knowledge acquired by the subjects is represented. Perhaps, as implied by the original artificial grammar studies, the subject acquires a whole system of If-Then productions and this procedural knowledge, like all procedural knowledge, is inaccessible to conscious introspection. On the other hand, perhaps the knowledge is not procedural at all, but declarative in nature. For example, subjects might abstract from the learning trials a prototype of a grammatical string; alternatively, they may simply memorize the instances on the study list. In either case, they may make relatively accurate grammaticality judgments by consciously comparing test items to the summary prototype, or to the specific exemplars they have memorized. We just do not yet know the answer. But what we do know is that amnesic patients can learn from their experience without remembering the learning experience itself, and in that sense, at least, implicit learning gives evidence of unconscious influence.

Implicit Thought

If the concept of implicit learning is more controversial than those of implicit memory or implicit perception, the concept of implicit thought is even more so. Still, the literature contains some favorable evidence, if only just a little. For example, some years ago Kenneth Bowers and his associates found that subjects can choose which of two problems is soluble without knowing the answer to the soluble one (Bowers, Regehr, Balthazard, & Parker, 1990). In some of their research, Bowers et al. employed a modification of the Remote Associates Test (RAT) (Mednick, 1962), used in the study of creativity, which they called "Dyads of Triads" (DOT).

In the RAT, subjects are presented with three words, such as:

LICK
SPRINKLE
MINES

and are asked to respond with a fourth word which is associated with each of the previous three (in this case,SALT).

In the DOT, subjects are presented with two RAT-like word sets, such as:

PLAYING

CREDIT

REPORT

STILL

PAGES

MUSIC

One triad coheres on a common (if remote) associate (i.e.,CARD), while the other one (barring psychotically loose associations!) does not. The subjects are asked to generate the associate for the coherent set; failing that, they are to guess which triad is the coherent one. Even when they could not say what the solution was, across five studies Bowers et al. found that subjects correctly chose the coherent triad about 58% of the time -- significantly better than would be expected by chance. They obtained similar results on other experiments, where subjects were presented with picture fragments, and asked to indicate which set could be reassembled to form a coherent image.

The above-chance "guessing" performance of Bowers' subjects seems to be a kind of semantic priming effect. Reading each of the stimulus items will prime the semantically related target item in memory, just as doctor primed nurse in Marcel's (1983b) lexical decision task. Presumably, if the amount of priming crossed some threshold of activation, subjects could report the solution to the problem. Presumably, there is a kind of subthreshold, preconscious level of activation which, while insufficient for conscious awareness, is enough to push the subjects' choice behavior in the right direction. But what kind of priming is it? Because the stimulus items remain clearly visible while the subject tries to solve the problem, memory is irrelevant, so the priming effect cannot be classified as implicit memory. For the same reason, it cannot count as implicit perception, either. Nor can it be classified as implicit learning: no items are repeated, and there is no feedback telling the subjects whether their choices are right.

In fact, a doctoral dissertation by Victor Shamas (nee Shames) found that soluble RAT items primed their solution words for lexical decision, even when the subject failed to produce the solution itself.

In the Bowers et al. experiment, what is being primed is the solution to a problem. The mental representation in question is neither a percept nor a memory, but rather an idea or an image -- in the broadest sense, it is a thought. By analogy with implicit memory, my colleagues and I have defined implicit thought as the influence of some cognitive representation, itself neither a percept nor an episodic memory, on experience, thought or action, in the absence of conscious awareness of that representation (Dorfman, Shames, & Kihlstrom, 1996; Kihlstrom, Shames, & Dorfman, 1996). Implicit thought may well underlie some of the most interesting facets of creative thought. In my view, intuition reflects a priming-based "feeling of knowing" similar to what we commonly see in studies of memory (Hart, 1965; Nelson, Gerler, & Narens, 1984); incubation reflects the gradual accumulation of strength of this primed idea; and insight reflects the emergence of the preconscious idea into the full daylight of consciousness.

Bowers' subjects were able to guess which triads are coherent and which were incoherent, even when they were unaware of the solution to the coherent set.  So the target of the coherent triad, the solution of the problem, is influencing the subject's behavior even though they're not consciously aware of what the solution is -- even though they're not consciously aware of the target.  Now the interesting thing about this is that the target is not a percept -- it's not as if the word “card” is being presented subliminally or in a masked fashion.  And it's not a memory either: it's not that the subject saw the word “card” earlier in the experiment and later forgot all about it.  Rather, in this experiment it’s the idea of “card” that's influencing the subject's performance on the Dyads of Triads.  It isn't a percept, and it's not a memory; it's an idea, or a thought, that's self-generated.  But it's a thought of which they're not conscious.  It's an implicit thought, not an explicit, conscious thought.

The idea of unconscious thought -- the idea that implicit thoughts can be dissociated from explicit, conscious thoughts -- takes a little while to get your head around, I'll admit.  But there's increasing evidence for the idea that implicit and explicit thought can be dissociated.  

  • First, there are experiments modeled after the Dyads of Triads paradigm, including evidence that presentation of a coherent triad can prime the processing of the target, even though the subjects aren't aware of what the target is. 
  •  Certain neurological patients with damage to areas of the prefrontal cortex can respond differently to risky choices, as opposed to safe choices -- risky choices meaning that there's a low probability of success, safe choices where there's a high probability of success -- even though they're not consciously aware of the probabilities involved.  Again, they seem to have a representation of the probabilities, but they're not consciously aware of that representation. 
  • And children can show something like implicit thought in experiments on insight learning.  If you give young children a series of problems, all of which can be solved the same way, you can show by decreases in the amount of time it takes them to solve each successive problem, that they've got it -- even though, when you ask them, they'll tell you that they don't know what the strategy is.  They're using a strategy to solve problems that they don't know they're using, because they don't know they have the strategy. 

The implicit thoughts revealed by Bowers' experiments with the Dyads of Triads procedure seem to have a lot in common with intuition -- those times when something just rings a bell, or just seems right, even though we can't say exactly why that's the case.  These intuitive feelings, in turn, seem to be related to the stages of thought described by Graham Wallas.  Preparation includes the accumulation of knowledge and the mastery of the rules which govern the particular domain in which some problem resides.  And it also includes the adoption of a problem-solving attitude, including the awareness that there is a problem to be solved and the deliberate, conscious analysis of the problem itself.  If it seems the problem can be solved with a well-known algorithm, then the problem just may be solved at this point and the thinker moves straight to the verification stage, which I'll discuss later.  Otherwise, the thinker proceeds to work actively on the problem. 

But if the problem is very difficult, the thinker may give up, or at least take a break, and here's where things get interesting.  Wallas proposed that the preparation stage is followed by a stage he called incubation, in which the thinker is no longer consciously thinking about the problem -- he's turned his attention somewhere else, or even gone to bed.  But in Wallas' view, problem-solving activity goes on anyway, outside of conscious awareness.  If the problem-solver is lucky, incubation is followed by intimation, a state of "rising consciousness", which indicates that a fully conscious flash of success is coming.  This intimation stage is otherwise known as intuition.  Then there's the illumination stage, in which the solution to the problem appears in the consciousness of the thinker.  This is a stage that we might want to call insight.  And finally, the verification stage in which the solution -- however it's been arrived at -- is confirmed and refined or shown to be incorrect -- in which case the problem-solver returns to the preparation stage.

In 2013, Errol Morris, a documentary filmmaker released a new film, The Known Unknown, about Donald Rumsfeld, who had been secretary of defense in the George W. Bush ("Bush II" or "Bush 43") administration, and vigorous proponent of the 2003 war in Iraq.  In a 2004 memo responding to questions about the existence of Iraq's weapons of mass destruction, Rumsfeld had famously noted that "There are known knowns.  There are known unknowns.  there are unknown unknowns."  In an interview in the film, Rumsfeld completed the implied 2-by-2 table.  As recounted by Maureen Dowd ("A Mad Tea Party", New York Times, 10/13/2013):
There are also unknown knowns.  Things that you possibly may know that you don't know you know.Morris challenges him: "But the memo doesn't say that.  It says that we know less, not more, than we think we do."Rumsfeld finally admits to a boo-boo: "Yeah, I think that memo is backwards."
Actually, Morris has it wrong.  He thinks that the concept of unknown knowns means that you know things that you really don't.  But it doesn't.  The concept of unknown knowns means, precisely, that there are things that we don't know that we actually know.  This is exactly the idea behind the notion of unconscious thought.  There are ideas which we have, and which influence our experience, thought, and action, but of which we are not consciously aware.

Incubation or unconscious problem-solving plays an important role in the folklore of problem-solving, but it's also one of the most controversial topics in the literature on thinking and reasoning.  There are a lot of instances where scientists and artists of various kinds have claimed to be confronted by some problem that appears to be insoluble.  They set the problem aside for a while, and then all of a sudden the solution to the problem appears in a flash.  It's as if they've been working on it unconsciously all that time.  On the other hand, many researchers and theorists deny that it occurs at all, and so far research hasn't decisively settled the issue.  In principle however, incubation is the stage in which the solution to a problem is coming into consciousness, but it hasn't yet arrived. 

But intuition appears to be what's going on in Bowers's Dyads of Triads task.  The solution to the coherent triad, the target of the problem, has not yet risen to the level of conscious awareness.  But unconsciously it influences the subject's ability to discriminate between coherent and incoherent triads, as well as the subject's performance on other kinds of tasks.  Intuitions arise in the period of incubation, and persist there as well, until they emerge into consciousness in the form of a full-blown insight.  Now these intuitive feelings are also controversial, and some theorists argue that they're misleading more often than not, and other theorists believe that too much conscious thought can actually make for bad decision making, and that people are better relying on their automatically generated unconscious intuitions.  But in Wallas' stages of thought, we can see two different aspects of unconscious thinking.  Incubation, where it appears that thinking itself goes on outside of conscious awareness, but the solution is not yet accessible to consciousness; and intuition, where the solution influences problem-solving behavior even though the solution has not yet appeared in the form of a full-blown conscious insight.


Where Does Hypnosis Fit?

The fourth line of research contributing to the rediscovery of the unconscious was hypnosis, a social interaction in which the subject acts on suggestions for experiences involving alterations in perception, memory, and the voluntary control of action. In fact, a case can be made that contemporary interest in unconscious processing arose in response to a single book,Divided Consciousness: Multiple Controls in Human Thought and Action, which took hypnosis as its major focus (Hilgard, 1977). Almost a century earlier, psychologists like William James (James, 1890/1980) had recognized that many of the phenomena of hypnosis involved a division in consciousness such that memories, percepts, and the like influence experience, thought, and action outside of phenomenal awareness. Hilgard put a modern spin on this earlier literature by reporting a series of carefully controlled experiments on suggested analgesia and other phenomena of hypnosis, and putting the research in the framework of modern cognitive theory.

For example, one of the earliest demonstrations of the dissociation between explicit and implicit memory occurred in a line of research on posthypnotic amnesia B the phenomenon which gave hypnosis its very name (J. F. Kihlstrom, 1985). After receiving appropriate suggestions, some individuals come out of hypnosis unable to remember the events and experiences which transpired while they were hypnotized -- an experience similar to the difficulty we all have, upon awakening, in remembering dreams and other events that transpired while we were asleep. People differ in their responsiveness to hypnotic suggestions, however, and posthypnotic amnesia is much more common in those who are highly hypnotizable (about 10-15% of the population), compared to those who are moderately hypnotizable or insusceptible to hypnosis.

One experiment classified subjects as low, medium, high, or very high in hypnotizability, on the basis of standardized tests devised for this purpose (J.F. Kihlstrom, 1980, Experiment 1). Regardless of their test scores, all subjects were treated as if they were hypnotizable. Following a standard induction of hypnosis, they memorized a list of 15 words; then they received a suggestion that, after coming out of hypnosis, they would be unable to remember what they did while they were hypnotized (posthypnotic amnesia does not occur unless it has been suggested by the hypnotist). When asked to recall the words they had memorized, the insusceptible subjects were able to do so readily, but the most highly hypnotizable subjects were densely amnesic, remembering less than a single word on average. After the amnesia suggestion was canceled, full recall was restored for all subjects.

While the amnesia suggestion was in force, the subjects received a free association test, in which they were asked to report the first three words that came to mind in response to cue words read aloud by the experimenter. The stimulus words were of two types:critical cues were semantically associated with words that the subjects had studied during hypnosis (e.g., the cue bread was associated with the studied item butter), while neutral cues were associated with target words that were entirely new (e.g., the cue ocean was associated with the target water, which had not been memorized). The subjects were more likely to respond with critical than neutral targets, providing another example of semantic priming. The most important finding, however, was that the magnitude of priming was the same in hypnotizable subjects, who did not remember having memorized the critical targets, as it was in insusceptible subjects, who remembered the learning experience clearly. Preserved priming shows that posthypnotic amnesia effects explicit memory, but spares implicit memory.

Hypnotic suggestion can also affect perceptual functions. When given hypnotic suggestions for blindness, for example, many hypnotizable subjects have the subjectively compelling experience that they no longer can see. However, Bryant and McConkey (1989) showed hypnotized subjects cards on which were printed a homophone (a word which has the same sound as another word but a different spelling, such as pain and pane) together with a disambiguating word (e.g.,body or window). Other cards were shown to subjects when they were not hypnotically blind, and still other cards were not presented at all. The subjects did not see the cards shown during blindness, and they were also unable to recall the words on these cards posthypnotically. However, on a subsequent spelling test they strongly tended to spell the homophones in accordance with the disambiguating context in which the words had been presented on the cards. This is another kind of semantic priming effect (Eich, 1984), because spelling performance is influenced by the semantic relation between the target word and its context. The fact that semantic priming occurred on the spelling test indicates that the words in question were processed by the subjects, despite the fact that they could not see them during hypnotic blindness.

Yet a third example of unconscious processing in hypnosis is provided by posthypnotic suggestion ‑- the phenomenon that, so our mythology tells us, gave Freud his first good insight into the psychological unconscious. In some sense, posthypnotic suggestion is a special case of implicit memory, because subjects act on a suggestion given during hypnosis, despite the fact that they cannot remember the suggestion itself. But posthypnotic suggestion is interesting in another way, because it is typically experienced as an involuntary, quasi-automatic response to the cue. In a doctoral dissertation by Irene Hoyt, hypnotized subjects were given a posthypnotic suggestion to press a key whenever a particular number appeared on a computer screen. After hypnosis was terminated, they were also instructed to press the key in response to a different number. On some test trials, the posthypnotic suggestion and the waking instruction were put into conflict. The posthypnotic suggestion did not always win out, once again shattering the myth of the coercive power of hypnosis.

But the most important finding was that even when the two did not conflict, response to the posthypnotic suggestion interfered with response to the waking instruction, and vice-versa. In other words, despite the experience and appearance of automaticity, processing the posthypnotic suggestion consumed a lot of cognitive resources. Posthypnotic response is unconscious, in the sense that subjects are unaware of why they behave as they do, or even that they are behaving as they are; but it has had no opportunity to be automatized by extensive practice.

In some ways, the unconscious influences observed in posthypnotic amnesia and hypnotic blindness resemble those observed in the amnesic syndrome and blindsight, but there are important differences as well. In the first place, the impairments produced by brain damage are permanent, while the phenomena of hypnosis are reversible. But more important, hypnotized subjects are in no sense brain damaged. It might be possible to argue, on the basis of the amnesic syndrome and blindsight, that the basis of conscious recollection is in the hippocampus, and the basis of conscious vision is in the striate cortex. But in hypnotic subjects there is no question of damage to these areas. This means that our explanation of the distinction between conscious and unconscious mental life will have to do more than appeal to the involvement of particular brain centers.

Of course, the subjects in subliminal perception experiments are not brain-damaged either. But again there is a difference: just as hypnotic subjects are in no sense brain-damaged, the stimuli in hypnosis experiments are in no sense subliminal. In subliminal perception, by definition, the stimuli are severely degraded by the conditions of their presentation -- in Marcel's (1983b) case, by a masking stimulus presented, on average, less than 1/10 of a second after the prime. Similarly, studies of implicit memory in normal subjects typically involve degraded encoding conditions or extremely long retention intervals (Jacoby & Dallas, 1981); (Nelson, 1978). In posthypnotic amnesia, by contrast, the items in question are deeply encoded and and well retained (as indicated by the reversibility of the amnesia), and in hypnotic blindness the items are clearly perceptible to everyone but the subject. For these reasons, hypnosis offers a different perspective on unconscious mental life than that afforded by brain damage or subliminal stimulation.

Moreover, response to the hypnotist's suggestions often seems, both to the subject and to an outside observer, to be involuntary and almost compulsive in nature. For this reason the phenomena of hypnosis might seem to be variants on automaticity, but hypnotic behavior is by no means automatic. In cognitive psychology, automatic processes are usually held to be independent of intention, consume no attentional resources, and leave no trace of themselves in memory. But nothing happens in hypnosis unless the subject is actively engaged in what it going on; it is easy to demonstrate that response to hypnotic suggestions consumes attentional resources; and hypnotized individuals retain full awareness of what they have done, posthypnotically, unless they have received a specific suggestion for posthypnotic amnesia. Much as the amnesic syndrome, blindsight, and subliminal perception show that mental contents as well as processes can be unconscious, so hypnosis shows that processes can be unconscious even though they have not been automatized. Again, hypnosis is of interest because it affords a unique perspective on unconscious mental life.



Implicit Cognition

So there's our first view of unconscious mental life.  What I have called the cognitive unconscious consists, in the first place, of automatic processes evoked in the course of perceiving and learning, remembering and thinking.  These automatic processes are unconscious because they operate outside of our conscious awareness, and also outside of our conscious control.  And second, we have documented dissociations between explicit and implicit perception, memory, learning, and thought.  Implicit percepts influence our experience, thought, and action in the absence of conscious awareness of the stimuli in the environment.  Implicit memories influence our experience, thought, and action, in the absence of conscious recollection of some event in the past.  In implicit learning, we employ knowledge acquired through experience without being consciously aware of what that knowledge is.  And in implicit thought, our experience, thought, and action, is influenced by ideas, which are not themselves either percepts or memories, even though we're not aware of what these ideas are.

"To have ideas, and yet not be conscious of them.  There seems to be a contradiction in that, for how can we know that we have them if we are not conscious of them?"  That's what Kant wrote at the end of the 18th century.  He also wrote, "Nevertheless we may become aware indirectly that we have an idea, even though we are not directly cognizant of the same."  That indirect influence comes in the form of priming effects and other effects that reveal implicit perception, memory, learning, and thought. 

Beyond the Cognitive Unconscious

Along with automaticity, implicit memory, implicit perception, implicit learning, and implicit thought comprise the cognitive unconscious (Kihlstrom, 1987). But cognition is not all there is to mental life: there is emotion and motivation as well (Hilgard, 1980). Accordingly, we are led to ask whether there is an affective unconscious, involving emotional feeling states, and a conative unconscious, involving motivational drives and goals, as well. Of course, emotional and motivational states may arise automatically, and in that sense result from unconscious processes. Many of the applications of automaticity to personality, social, and clinical psychology make just this argument (e.g., Bargh, 1990; Zajonc, 1980, 1984). But can emotions and motives be unconscious, in the same way that memories can, and still dynamically influence our behavior? We are admittedly verging near Freudian territory here, but once one has accepted the evidence for implicit memory and other aspects of the cognitive unconscious, one must at least entertain the possibility that other mental states, including noncognitive states such as feelings and desires, can also be unconscious.

Of course, putting the question this way assumes that emotions and motivations are different from cognitions. There is a school of thought within cognitive psychology and cognitive science that denies that this is the case (we might call this view cognitive hegemony). Put bluntly, this school, whose roots go back to Aristotle, holds that our feelings and desires are cognitive constructions --beliefs about what we feel (Schachter & Singer, 1962) and what we desire (Lepper, Greene, & Nisbett, 1973). If emotional and motivational states are simply variants on cognitive states, and cognitive states can be unconscious (as in the case of implicit memories), then it follows logically that emotional and motivational states can be unconscious as well. On the contrary, if -- as Immanuel Kant argued -- "there are three irreducible faculties of mind: knowledge, feeling, and desire", and emotional and motivational states are not reducible to cognitive states, then questions about unconscious emotion and motivation must be answered on their own grounds.

Certainly our feelings and desires can be generated by unconscious, automatic processes, but in that case we'd still be aware of what we feel and of what we desire.  As in the case of unconscious cognition, the really interesting question is whether we can have feelings and desires the way we have ideas, and yet not be conscious of them.  Work on unconscious emotion and unconscious motivation is not nearly as advanced as work on unconscious cognition.  But still, there are reasons for thinking that we have unconscious feelings and unconscious desires.  The same way we have unconscious ideas.

The Motivational Unconscious

On the motivational side,the late David McClelland and his associates have articulated a concept of implicit motives -- interestingly, without overt reference to the concept of implicit memory (McClelland, 1980; McClelland, Koestner, & Weinberger, 1989). In parallel with the now-familiar distinction between explicit and implicit memory, explicit motivation may be defined as the conscious representation of a conative state, or the desire to engage in some particular activity, as represented by a craving for food, yearning for love, and the like. By contrast, implicit motivation refers to any change in experience, thought, or action that is attributable to one's motivational state, in the absence of conscious awareness of that state.

For McClelland and his colleagues, explicit motives are self-attributed: the person is aware of the motive, can reflect on it, and can report its presence in interviews or on personality questionnaires.  For example, the need Achievement subscale of the Personality Research Form (PRF) a widely used multivariate personality inventory, includes such items as "I enjoy doing things which challenge me" or "I will keep working on a problem after others have given up". Subjects' responses on such tests reflect their awareness of what they want, what their goals are, and what drives them.  If somebody endorses questionnaire items like "I enjoy the challenge of a difficult task", or "I maintain high standards in everything I do", or "I have a strong desire to excel", we can be pretty sure that they have a high level of achievement motivation.  And we can be pretty sure that they know it, they're consciously aware of it. 

Implicit motives, by contrast, may be inferred from the person's performance on such tasks as the Thematic Apperception Test (TAT), a classical "projective" test of personality. In the TAT, subjects are asked to make up stores in response to ambiguous pictures, and these stories are coded for various themes. For example, in response to a picture of a boy staring at a violin, one subject might talk about how the boy is being forced to practice the violin, when in fact he would rather be out playing baseball; another subject might say that he is suffering stage fright before a school performance in front of his parents; another that he is a child prodigy, supremely confident, about to premiere a new concerto before the crowned heads of Europe. Is the acrobat afraid?  Or is he about to perform confidently before the crowned heads of Europe?  In theory, the motives attributed to the characters in the picture are unconscious projections of the subject's own motives.

We are admittedly verging near Freudian territory here, but the motives in question are not seething sexual and aggressive impulses arising from the id; they are standard social motives for achievement, power, affiliation, and intimacy.

One thing we know for sure is that assessments of motives by means of the TAT do not relate strongly to assessments of the same motives by means of questionnaires.  Ordinarily, we'd expect two tests assessing the same personality trait, like the need for achievement, to correlate highly with each other -- correlation coefficients in the 60's or 70's and even 80's.  By contrast, the correlation between TAT measures of achievement motivation and questionnaire measures of achievement motivation are near zero.  The implication is that someone can score high on an implicit test assessment of achievement motivation, but score low on a questionnaire measure.  In that case, we might want to say the person really is motivated to achieve, he just doesn't know it.  He just doesn't consciously present himself that way, or perceive himself that way.

Similar findings have been obtained with respect to the need for power or dominance, and also the need for affiliation.  The fact that ostensible measures of explicit and implicit motives are so poorly correlated with each other doesn't by itself prove anything.  But it does suggest that people's behavior may be driven by motives of which they themselves are not aware.

On the basis of an extensive program of empirical research, McClelland and his associates concluded that explicit and implicit motives differ in a number of ways.

  • Explicit motives are strongly tied to prevailing social norms, while implicit motives are more likely to diverge from what significant others say we should desire.
  • Explicit motives are correlated with the choices that people make over the short run, while implicit motives are associated with long-term behavioral trends.
  • Explicit motives are aroused by extrinsic social demands, while implicit motives are aroused by intrinsic task incentives.

These differences are analogous to the dissociations found between explicit and implicit memories, and they suggest that we can be engaged in motivated, goal-oriented behavior without being aware of what our motives are.

Still, there are some serious methodological problems to be solved before we can accept this conclusion with confidence. The means by which explicit and implicit motives are assessed are so different that we cannot be confident that the fundamental distinction has to do with awareness. For example, the PRF and TAT may assess different motives, even though their scales have the same names. If so, it would not be surprising that they correlated with different classes of behavior. Moreover, it has not been demonstrated that people are unaware of their motives, as assessed by the TAT. The subject who projects high achievement motivation on a boy in a picture may be perfectly well aware of his own achievement motivation. Proper documentation of a dissociation between explicit and implicit motivation requires that subjects be unaware of the motives which implicitly influence their experience, thought, and action. The literature on implicit motivation rarely asks this question, and it must be satisfactorily addressed before we know what to make of implicit motives.

The Emotional Unconscious

Turning next to the affective domain, my collaborators and I have proposed that we seek dissociations between explicit and implicit emotions (Kihlstrom, Mulvaney, Tobias, & Tobis, 2000). Again paralleling the vocabulary of the cognitive unconscious, we define explicit emotion as the person's conscious awareness of an emotion, feeling, or mood state; implicit emotion, then, refers to changes in experience, thought, or action which are attributable to one's emotional state, in the absence of conscious awareness of that state.

Emotion as an Expression of Implicit Cognition

Of course, it is well known that conscious emotional responses can serve as expressions of implicit memory and perception, Most of the relevant studies make use of the mere exposure (Zajonc, 1968), in which prior exposure to an object increases one's preference for that object on a subsequent choice task. According to one class of theories, the mere exposure effect is a result of the repetition priming effect already familiar from studies of implicit memory and implicit perception. The initial exposure creates, or activates, representations of the stimulus in memory. When on a later occasion the person has to make a preference judgment about the same object, the resulting memory-based feeling of familiarity increases likability. There is a similar repetition effect on judgments of truth (Schwartz, 1982), and the underlying mechanism is probably the same.

On the memory side, it seems that brain-damaged, amnesic patients show the mere exposure effect, even though they cannot consciously remember the objects to which they had previously been exposed (Johnson, Kim, & Risse, 1985). And in the study that really got the affective revolution going (Zajonc, 1980), Kunst-Wilson and Zajonc (1980) found that normal subjects showed mere exposure effects even when the exposures in question were subliminal, and thus not consciously perceived. Subsequent research has indicated that the subliminal mere exposure effect may be even more powerful than the supraliminal one (R. F. Bornstein, 1989, 1992). In both cases, the subjects knew what they liked, but did not know why. Their feelings are a little reminiscent of those expressed by the 17th-century English satirical poet, Thomas Brown:

I do not love thee, Dr. Fell.
Why I cannot tell.
I only know
I do not love thee, Dr. Fell.

The idea that emotional responses can reflect implicit memory is at least as old as Breuer and Freud (1893-1895/1953), whose hysterics "suffered from reminiscences" of previous traumatic events. More recently it has been revived as part of what has been called the trauma-memory argument -- that memory traces of traumatic events, inaccessible to conscious recollection, may nonetheless influence the victim's experience, thought, and action in the form of pathological symptoms (Kihlstrom, 1995, 1996b, 1997, 1998a). For example, Frederickson has distinguished between consciously accessible recall memories and unconscious feeling memories of trauma (Fredrickson, 1992). As she writes, "feeling memory is often experienced as a flood of inexplicable emotion.... These clients are experiencing a feeling memory about being abused, even though... they can recall nothing about their abuse" (p. 92).

The notion of "feeling memories" finds some support from the literature on emotion as an expression of implicit memory, but some clinicians erroneously infer a history of trauma from their patients' reports of inexplicable feelings, and then engage in therapeutic practices, sometimes called recovered memory therapy, designed to restore traumatic memories to conscious recollection (Kihlstrom, 1996b, 1997, 1998a; Shobe & Kihlstrom, 1997). Unfortunately, a link between some past event and a person's current emotional state can be made only when there is independent corroboration of the event in question (Kihlstrom, 1997). When there is no independent corroboration of the traumatic event, however, clinical inferences of past trauma violate the logic of implicit memory, and can lead therapist and patient onto the dangerous ground of false memory syndrome (Kihlstrom, 1998a).

Implicit Emotion

But what of emotional states themselves?  Can they be unconscious?  Some theorists assert that the idea of unconscious emotion is a contradiction in terms (e.g., Clore, 1994). But there are reasons to take the idea of unconscious emotion seriously.

Again by analogy to implicit memory, we can define explicit emotion as the person's subjective mood state -- subjective, affective response or feeling state; and implicit emotion as any effect on the person's experience, thought, or action that's attributable to the individual's emotional state -- even in the absence of conscious awareness of that emotional state.  And again, these effects are most likely to be seen either in terms of the person's overt motor responses, like facial expressions or, covert physiological responses, such as might be measured in the autonomic nervous system. 

Peter Lang's multiple-systems theory of emotion suggests that we may be able to dissociate implicit unconscious emotion from explicit conscious emotion.  Recall that theory asserts that there are three different components to every emotional response: a verbal-cognitive component, representing the subjective experience of emotion; overt motor behavior; and covert physiological activity.  These three components or systems usually covary, but under some circumstances they can move in different directions -- a state which Rachman and Hodgson have labeled desynchrony (Rachman & Hodgson, 1974). Desynchrony comes in several forms. We sometimes see this in the treatment of phobias, where after treatment a patient may say that he’s not afraid of something he used to be afraid of, but he’ll still avoid it, and his heart will start pounding when he encounters it. The desynchrony that is relevant in the present context occurs when explicit emotion, as represented by a person's conscious, subjective feeling state, is absent, but the person nevertheless shows recognizable behavioral and somatic signs of implicit emotion.

Empirical evidence for this particular form of desynchrony is sparse, and rarely goes beyond clinical anecdote (Kihlstrom et al., 2000), but at least it is predicted by a well-established neuropsychological model of fear proposed by LeDoux (1996). LeDoux proposes that environmental stimuli are first processed by sensory centers in the thalamus, which then pass information about emotional events to the amygdala, which in turn generates appropriate behavioral, autonomic, and endocrine responses. Information about these responses is also passed to cortical centers supporting working memory, where it is integrated with information provided by thalamic centers about the fear stimulus, thus generating the full-blown subjective experience of being afraid of something.

However, a disconnection between thalamus and cortex will prevent the fear-eliciting stimulus from being represented in working memory. As a result, the person will experience fear without being aware of the fear stimulus. In this case, emotion will serve as an implicit expression of perception or memory, as described earlier.

Alternatively, if there is a disconnection between the amygdala and the cerebral cortex, the person will behave in a fearful manner without feeling fear or anxiety. In this case, there will be a dissociation between explicit and implicit emotion.

Ordinarily we think of these three components

To summarize the argument so far: Lang's multiple systems theory of emotion predicts that there might be dissociations between conscious and unconscious emotion.  And Joseph LeDoux's neuroscientific theory of emotion suggests that such dissociations could in fact happen.  If, for example, the brain module that mediates the verbal-cognitive component of emotion is damaged somehow, but the brain modules that mediate the overt and covert components of emotion remain intact.  

So it can all work out in theory, but it would be nice to have a little bit of evidence.  Some evidence along these lines comes from research on social attitudes – how whites feel about blacks, how men feel about women, and vice-versa.  In traditional social-psychological research, attitudes are defined as affective dispositions to like or dislike certain things, and they are usually measured explicitly by self-report scales. In other words, the traditional concept of attitude is of a conscious feeling state -- an explicit emotion. However, Greenwald and Banaji (1995) have suggested that people may possess positive and negative implicit attitudes about themselves and other people, which can also affect ongoing social behavior outside of conscious awareness. Their Implicit Association Test (IAT) employs reaction time to measure the amount of time it takes to judge whether something is good or bad.  A number of studies have shown that implicit attitudes, so measured, do not correlate very highly with explicit attitudes as measured by self-report.  Unfortunately, the experimental literature on implicit attitudes rarely offers a direct contrast with between explicit attitudes, so we don't yet know whether such dissociations actually occur, in what kinds of people, and under what circumstances. Still this experimental approach to implicit emotion is very promising. So it might be that the attitudes tapped by the IAT are truly unconscious; or it might be that the subjects simply were reluctant to divulge their true attitudes on the explicit measure, isn’t completely clear yet – but it’s a start. 

The Psychological Unconscious Now


Although the research I've just described gives us sufficient reason for taking unconscious mental life seriously, it has to be said that the evidence in each of these domains is not equally strong.  I think that the evidence for implicit perception and implicit memory is quite convincing, establishing the existence of unconscious percepts and memories to the satisfaction of all but very few critics.  Implicit learning is also very widely accepted, though a little more controversial.  Evidence for implicit thought is on somewhat softer ground -- but only because there have been relatively few studies in this domain compared to perception, memory and learning.

The same holds true for unconscious motives and unconscious emotions.  Research in these domains is just getting off the ground. 

Earlier in the Principles, and perhaps with Hartmann in mind, James had offered a warning which would reverberate throughout the 20th century exploration of the unconscious:

The distinction between the unconscious and the conscious being of the mental state is the sovereign means for believing what one likes in psychology, and of turning what might become a science into a tumbling-ground for whimsies (1890/1980, Vol. 1, pp. 163).

It's all too easy for us, as psychologists, or even as ordinary people on the street, to tell people what they unconsciously believe, or think, or feel, or want.  How could they possibly contradict us? 

James's criticism was prophetic, as we can see in the ensuing controversy, ranging across the entire 20th century, concerning Freud's psychoanalytic theory of mental life. Nevertheless, studies of automatically controlled processes, and of explicit and implicit perception, memory, learning, and the like, offer a solution to James' warning, because they offer strict criteria for identifying unconscious mental processes and unconscious mental states, and for making inferences about subjects' unconscious mental lives, by tying these inferences to objective evidence of behavior in the controlled environment of the laboratory.

  • For example, the scope of the psychological unconscious, broad as it is, does not appear to be so broad as to encompass the primitive, infantile, irrational unconscious of Freud's psychoanalytic theory. There is no evidence, in any of the science I have summarized today, favoring Freud's view that the unconscious is the repository of primitive, infantile, irrational, sexual and aggressive impulses, repressed in a defensive maneuver to avoid conflict and anxiety.
  • Nor is there evidence that subliminal influences have special power to influence what people think, feel, want, and do.
  • Nor is there any evidence to support clinical lore concerning unconscious representations of trauma, or the excesses of the recovered memory movement in psychotherapy. In this last case, as James warned, the unconscious does indeed seem to be a tumbling-ground for whimsies.

The rediscovery of the unconscious may have gotten a late start in psychology, but the study of unconscious mental life is now a growth industry in psychology. With a sophisticated, ever-enlarging understanding of cognitive processes, it is now possible to discuss unconscious mental life without making any substantive reference to Freud. Although there remain some skeptics, research on unconscious cognition now focuses less on existence proofs and more on analyses of its scope and limits. And research on unconscious emotion and motivation, modeled on studies of the cognitive unconscious, has begun to develop independent of psychoanalysis.

In fact, evidence of unconscious mental life is so vast, so convincing, increasing and strengthening so much with each new issue of our best journals, that we are in danger of coming full circle, to the position of conscious inessentialism discussed (but not endorsed) by the philosopher Owen Flanagan.  Some philosophers and psychologists have concluded that automaticity and priming dominate our behavior, so that consciousness is an illusion -- at best a commentary on what is going on down below, at worst a delusion that gets in the way of adaptive action (not to mention a proper scientific understanding of mind and behavior). I think we see this trend clearly in the "automaticity juggernaut" that I described earlier.

But that's a mistake too.  Unconscious mental life exists, but it's not the whole story.  There is still plenty of room for deliberate, conscious thought to influence both experience and action.

This page last revised 09/26/2014.