While philosophical analyses of consciousness began with Descartes, the scientific study of consciousness was delayed until the beginning of the 19th century and the emergence of scientific, as opposed to philosophical, psychology. The first problem for the new psychology was the nature of sensory experience -- that is, conscious sensory experience. In this way, scientific psychology began, quite expressly, as the scientific study of conscious experience.
The initial emphasis on sensation -- as opposed to memory or thinking or some other aspect of consciousness -- was quite deliberate, because the earliest psychologists understood that, in order to be scientific, they had to be able to tie conscious experience, which is inherently subjective, to the objective world of physical events. And so they began by exploring the relations between subjective sensory experiences -- what the philosophers call qualia -- and the objective physical stimuli that gave rise to them -- hence, the new field of psychophysics.
With respect to qualia, we can distinguish, along with Kant, qualitatively different mental states of knowing, feeling, and desiring. To paraphrase Nagel, there is something it is like to know something, and that "something" is different from what it is like to feel or desire something. The basic (and, if Kant is right, irreducible) mental faculties of knowledge, feeling, and desire are associated with distinct qualitative experiences of knowing, feeling, and desiring. If Searle is right, each of these superordinate qualia is associated with a different form of intentionality, but there's obviously more to qualia than that.
Even within the faculty of knowing,
Sensation is the result of some pattern of physical energy radiating from some object and impinging on an organism's sensory organs -- organs which are composed of specialized neural tissue, and which are connected to the central nervous system by afferent nerves.
How many senses are there? Writing in the 4th century BCE, Aristotle (in De Anima) famously identified five senses: vision (seeing), audition (hearing), olfaction (smelling), gustation (tasting), and touch (touching and feeling).
That's a good start (Aristotle is always a good start), but following Sherrington (1906), we can identify at least nine different sensory modalities, or general domains in which sensation occurs. These modalities may be arranged hierarchically (of course!), beginning with a division into three broad categories, exteroception, proprioception, and interoception. Within each of these categories, there are (of course!) further subcategories.
Notice, in passing, that many of the nine
senses are variations on two different mechanisms.
The organization of the sensory modalities illustrates the basic concept of transduction: the conversion of some physical stimulus energy into a neural impulse. Looking over the sensory modalities, it appears as if each sensory receptor is specifically responsive to a different proximal stimulus energy (which, in turn, radiates from a distal stimulus). In transduction, a particular type of physical energy is converted into a neural impulse, which is carried via a sensory tract to a particular part of the brain known as a sensory projection area.
Thus, in vision, the proximal stimulus consists of light waves (electromagnetic radiation) emitted by or reflected from an object. Not all light waves give rise to the experience of seeing, however; in humans, only wavelengths of 380-780 nanometers (nm, or billionths of a meter) are visible. Other wavelengths, such as infrared (greater than 780 nm) and ultraviolet (less than 380 nm) are not visible to the human eye.
In any event, light waves pass through the cornea
and lens of the eye, which inverts the image and
focuses the light on the retina on the back of the eyeball.
This is the retinal image. There, light waves impinge
on two types of receptor organs: rods and cones.
Farther along on their journey to the brain, the visual impulses pass through the lateral geniculate nucleus (LGN), a bundle of neurons which is part of the thalamus. Each cell in the LGN corresponds to a particular part of the retina; thus, the LGN processes information about the spatial organization of the visual image. With the primary exception of olfaction, most afferent impulses pass through the thalamus on the way to their sensory projection areas (olfactory impulses go directly to a part of the brain called the olfactory bulb). Thus, the thalamus acts as a kind of a sensory relay station, directing impulses representing various modalities to their appropriate cortical projection areas.
Finally, the visual impulses arrive at their sensory projection area, which is the visual cortex of the occipital lobe, known as V1, or Brodmann's area 17. Again, each cell in the visual cortex corresponds to a particular part of the retina, in what is known as retinotopic organization.
According to a common view of perception, each
sensory receptor organ is specific to a different kind of
proximal stimulus, resulting in the various sensory
Proximal Stimuli and Receptor Organs Associated with the Sensory Modalities
Wavelength: 380-780 Nanometers
|Rods and Cones on Retina|
Frequency: 20-20,000 Cycles per Second
|Hair Cells of Basilar Membrane|
|Chemical Molecules in
Dissolved in Saliva
|Papillae on Tongue Taste Buds|
Molecules in Air,
Dissolved in Mucous
|Olfactory Epithelium in Nasal Cavity|
Perifollicular (Basket) Endings
Inflammation of Injured Skin
Free Nerve Endings
Otoliths in Semicircular Canals
|Hair Cells in Vestibular Sac|
"Sensation consists in the sensorium's receiving... a knowledge of certain qualities.... of the nerves of sense themselves; and these qualities of the nerves of sense are in all different, the nerve of each having its own peculiar quality or energy."
According to Muller, the modality of sensation is not determined by the proximal stimulus alone, but rather by the stimulation of particular nerves. These nerves are ordinarily responsive to some specific proximal stimulus, but they can also respond to others. You can have a sensory experience without a proximal stimulus, so long as some particular receptor organ is stimulated. The implication is that the modality of sensation is not determined by the proximal stimulus, but by the sensory receptor organ that happens to be stimulated.
As it happens, this isn't quite true either.
In the first place, we now know, from the Nobel Prize-winning work of Lord Adrian, that the neural impulses generated by the different receptor organs are all the same -- they're electrical impulses known as action potentials. It's not the case that the rods and cones generate one kind of "nerve energy", and the hair cells in the cochlea generate another kind. All neural impulses are identical, in that sense. Just as all the wires in a radio carry the same electrical current, regardless of their size or composition or the color of their insulation, all neurons generate the same action potentials.
The proof comes from electrical stimulation of the afferent nerves leading from the receptor organ to the central nervous system, or of the sensory projection areas in the cerebral cortex. Either case produces sensory experience in some modality, vision or audition or whatever, even though there has been no proximal stimulus in the usual sense nor any activity in any sensory receptor per se.
So if it's not a particular proximal
stimulus, or even the activity of a particular receptor
organ, what determines the modality of sensation? In
the abstract, the modality of sensation must be determined
by one of four factors, or some combination of these four:
And in audition:
Sensory Tracts and Cortical Projection Areas Associated with the Sensory Modalities
| Optic Nerve (Cranial Nerve
Lateral Geniculate Nucleus
| Area V1 of Occipital
(Brodmann's Area 17)
|Cochlear Branch of Vestibulo-Cochlear
Nerve (Cranial Nerve VIII)
Medial Geniculate Nucleus
|Area A1 of Temporal Lobe
(Brodmann's Area 41)
|Sensory Portion of
Glossopharyngeal Nerve(Cranial Nerve IX)
Facial Nerve (Cranial Nerve VII)
Vagus Nerve (Cranial Nerve X)?
| Anterior Insula
Olfactory Nerve (Cranial Nerve I)
| Prepyriform Cortex
Some Cranial Nerves
| Somatosensory Cortex
Some Cranial Nerves
| Somatosensory Cortex
Spinal, Some Cranial Nerves
A-delta Fibers, Neo-Spinalthalamic Tract
C Fibers, Paleo-Spinalthalamic Tract
Spinal, Some Cranial Nerves
|Cranial Nerves||Parietal Lobe|
Pain is an interesting case, and I'll discuss it in more detail later.
Both types of pain signals end up in the primary somatosensory cortex -- at least for . However, the suffering com"suffering
Although in principle, the modality of sensation is determined by four factors (proximal stimulus, receptor organ, sensory tract, and sensory projection area), in the final analysis the projection area is the most important. If the auditory nerve were connected to the visual cortex, we would see sounds instead of hearing them (we know this because direct electrical stimulation of the visual cortex produces visual sensations). Regardless of where the sensory impulse comes from, what matters is where it goes. The Doctrine of Specific Nerve Energies is correct, but the nerve energies in question have more to do with the sensory projection areas than they do with the sensory receptors or the sensory tracts.
With respect to the four characteristic features of a sensory modality -- proximal stimulus, receptor organ, sensory tract, and projection area -- some modalities -- especially vision and audition -- are clearly organized and well understood. The other modalities are apparently more complicated; certainly they are more poorly understood.
For example, the chemical senses interact. (1) a head cold clogs the nasal passages, preventing the person from smelling properly; but food is also tasteless. (2) Cigarette smoke clogs the taste pores, but perfume is also odorless.
The skin senses are hopelessly complex.
At least six different nerve endings are found in the skin,
but there is no isomorphism (1:1 relation) between the type of
receptor, and the corresponding sensory experience when that
receptor is stimulated electrically. To make things worse,
neural impulses from all six receptor types are transmitted
along the same afferent tracts to the spinal cord and up to
the brain, and all project to the same somatosensory cortex in
the parietal lobe. So the puzzle is: how are the skin senses
kept separate? There is a prize for the person who comes up
with the best answer, which you can collect in Stockholm.
Consider, as well, the experience of pain -- a common example offered in philosophical discussions of qualia.
The stimulation of a single tooth results in the eventual activation of no less than five distinct brain-stem pathways.... Two of these pathways project to cortical somatosensory areas 1 and II, while the remainder activate the thalamic reticular formation and the limbic system, so that the input has access to neural systems involved in affective as well as sensory activities [references deleted].In the final analysis, the experience of pain is largely determined by the meaning of the stimulus. In doctors' offices, people flinch when there isn't much pain. The responses of injured children differ markedly depending on whether an audience is present. Placebos provide genuine relief of pain. Soldiers experienced different levels of pain from objectively similar wounds in World War II, Korea, and Vietnam. And many religious rituals in tribal cultures (e.g., fertility, puberty) would appear objectively painful, but elicit no signs of pain from the participants.
In the present context, however, the most important point is that "there's something it's like" to see something, and that is different from what it's like to hear something. After Kant's "three irreducible faculties", there appear to be nine qualitatively different forms of sensation -- seeing, hearing, tasting, smelling, feeling touch, feeling warm and cold, and feeling pain. And each of these modalities is associated with the activity of a particular neural system.
Psychology began as the scientific study of consciousness, even though it quickly lost that interest, and regained it only recently. At the beginning, psychologists relied on introspection -- literally, "looking into" the mind to see what is there.
The roots of late-19th-century introspection lie in the psychophysics of the early 19th century, which examined the relationship between the physical properties of a stimulus (subject to third-person objectivity) and the psychological properties of the corresponding experience to which that stimulus gives rise (subject to first-person subjectivity).
The psychophysicists were especially interested
in the intensity of stimulation, though they worked the same
way for other sensory qualities. Their essential method
was to vary the physical intensity of a stimulus, and then ask
the subject to rate, on a numerical scale, the intensity of
the corresponding sensory experience. By this means,
they focused their research on two different types of sensory
Of course, the absolute threshold is really just a special case of the relative threshold -- that is, the difference between "nothing" and "something".
Actually, there were three different
The Scoville Scale
An interesting application of classical psychophysics is the "Scoville Scale" that measures the "hotness" of chili peppers. In the Scoville scale, a standard amount of ground chili pepper is dissolved in a standard amount of syrup. The resulting ratio yields a measure of the "heat" in the pepper in question.
According to the psychophysical principle, every psychological property of a sensation is related to some physical properties of the corresponding stimulus.
In vision, for example, the quality of hue (or color) is related to the wavelength of electromagnetic radiation that falls on the retina of the eye. Short wavelengths correspond to blue, longer wavelengths to yellow, etc.
Another visual quality, saturation, is related to the amount of gray that is mixed with the color. For example, pink has more gray than red.
In audition, the quality of pitch is related to the frequency of sound waves falling on the eardrum. In music, middle C, between the treble and bass clefs, has a lower frequency than third-space C on the treble clef.
Another auditory quality, timbre, is related to the shape of the sound wave, which in turn is related to the number and distribution of harmonics above the fundamental frequency. The sound of a flute is characterized by a fairly "pure" sine wave, while the raspy sound of an oboe is characterized by a more square wave.
Let's take the simplest psychophysical experiment first, on the absolute threshold for sensation. The question is: what is the minimum amount of stimulus intensity which will give rise to a conscious sensation?
Note that the
shape of the actual curve relating intensity to detection
is smooth and ogival (that is, it takes the form
of an S-shaped curve). This fact, coupled with the existence
the very concept of the threshold. At the very
least, it seems that the threshold is a somewhat
variable border between what is consciously perceptible
and what is not consciously perceptible -- a
border that can be placed at somewhat different
locations depending the observer's biases and
incentives, as well as sensory acuity.
Such factors are taken into account by signal
detection theory, which provides separate estimates
of the sensitivity of the sensory system and the biases
of the observer.
But even in signal-detection theory, there is still some
notion of the threshold -- some level of stimulus
intensity that even the most sensitive system can't
What, besides the obvious factor of stimulus intensity, makes the difference between supra- and sub-threshold stimuli? Obviously, the answer must lie in the way in which the brain processes the stimulus. A study by Bram van Vugt and his colleagues may shed some light on this question (Science, 2018; image at left from a commentary on the article by George Manshour). They trained monkeys to report very weak visual stimuli (with a very low degree of contrast between the stimulus and its background). Now, of course, monkeys can't make verbal reports of their conscious experience, but they can be trained to show by their behavior (in this case, making a particular eye movement) whether they've detected a stimulus or not -- and that can be a proxy for conscious awareness. And monkeys have other advantages over humans: the investigators were able to implant electrodes in the occipital and prefrontal cortex to record multiunit brain activity in response to visual stimuli (not single units, but rather the activity in small clusters of adjacent neurons). They then presented the monkeys with visual stimuli varied across three levels of contrast, making them very difficult to detect (correct response < 40%, so essentially subthreshold or subliminal), moderately difficult (40% < correct response < 60%, or right around the limen or threshold), or relatively easy (> 60%, clearly supraliminal or suprathreshold).
These findings are
generally consistent with a version of Global Workspace
Theory (Baars, 2002) discussed in the lectures on the Neural Correlates of
Consciousness -- specifically, the Global Neuronal
Workspace Theory (GNWT) proposed by Dehaene &
Changeux (2011), which identifies GW with prefrontal
cortex. Apparently, conscious experience is a
product of two processes: propagation of a
signal through the brain, from the sensory cortices to
the prefrontal cortex; and ignition, in which
the prefrontal cortex broadcasts the signal to a wide
variety of cortical centers. It is this ignition
process that results in a conscious sensation.
Here's more detail:
Of course, the whole scheme depends on the equation of the
monkeys' behavior with conscious awareness and
reportability. That's not a bad assumption. But,
as we'll discuss later in the lectures on Explicit
and Implicit Cognition, that's not the only
Most early psychophysical research was focused
on the dimension of intensity, or the experienced "strength"
of a stimulus. In terms of the psychophysical laws, the
intensity of experience was related to the amount of physical
energy emanating from the stimulus and falling on the
observer's sensory surfaces. Intensity is coded in the
nervous system by the temporal and spatial summation of
activity in sensory neurons.
This much would seem fairly obvious. But, in fact, there is no isomorphism between physical and sensory intensity. For example, the relative threshold -- the amount of change in physical energy needed before that change is detectable -- depends on the intensity of the original stimulus. If the original intensity is low, even a very small change is noticeable. If it is high, a relatively large change is required.
This principle is represented in Weber's Law:
In other words, the amount of intensity which must be added to a stimulus to produce a JND is a constant fraction of that intensity. To take an example, assume that Weber's Fraction, C, = 1/10. Thus,
If the intensity of the original stimulus is:
Change is first noticeable at:
110 units, not 101
220 units, not 201
You may wish to do some other problems like this, using different Weber Fractions and different values of I, just to get a feel for the enterprise.
differs for each sensory modality, and for each species, and
provides a universal index of the sensitivity of a modality.
Thus, if a modality is associated with a small Weber's
Fraction, it is highly sensitive (small JNDs). In humans,
vision and audition are the most sensitive modalities, taste
and smell the least. But in cats, Weber's Fraction for smell
is 1/300 that of humans.
A more general principle relating sensory intensity to physical intensity is Fechner's Law:
Thus, for every value of I, a corresponding value of S is given by the formula. Ignore the constant k, and assume logarithms to the base 10 (you may wish to get out your calculator and work through some other examples). Thus,
If the physical intensity is:
Sensory intensity is:
0 units (log101 = 0)
1 unit (log1010 = 1)
2 units (log10100 = 2)
2.3 units (log10200 = 2.3)
Notice that according to Fechner's Law,
As stimulation grows from 1 to 200 units, sensation grows only from 0 to 2.3 units). This follows from Weber's Law: 10 additional units of stimulation makes a big difference to sensation at the bottom end of the scale, but a progressively smaller difference as we move towards the top.
If you re-plot the
relation between S and I on a logarithmic scale of I, you get
a straight line. The 19th-century psychophysicists loved
findings like this, because the simplicity and elegance of the
straight line suggested that they had uncovered a basic law of
mental life after all.
Fechner's Law holds across a wide range of
situations, but there are some exceptions. For example:
An even more general psychophysical law, Stevens' Law, handles the exceptions. In fact, Stevens' Law is so general that handles all known psychophysical relations. Stevens' Law states that:
S=k(IN), where S = sensory intensity;
I = physical intensity;
k = a constant; and
N = an exponent (either power or root).
In words, Stevens' Law states that
for any quality of sensation, there is an exponent that relates changes in sensation to changes in stimulation.
Get out your calculators again, and we'll work through some examples.
In fact, Fechner's Law is a special case of Stevens' Law, where the exponent, N, is less than 1.0 (an exponent of 1/2 means the square root; an exponent of 1/3 means the cube root, etc.). As an example, consider what happens when N = 1/2 (or, alternatively, 0.5):
If the physical intensity is:
Sensory intensity is:
1 units (11/2 = 1)
3.16 units (101/2 = 3.16)
10 units (1001/2 = 10)
14.14 units (2001/2 = 14.14)
Again, although the precise values differ from those given in the illustration of Fechner's Law, notice that big changes in I yield small changes in S, just as predicted by Fechner.
Where N = 1, you get the "length" exception to Fechner's Law, where every change in I produces an isomorphic change in S.
Where N is greater than 1.0, you get the "pain" exception. For example, assume that the exponent N = 3/2 (or 1.5:
If the physical intensity is:
Sensory intensity is:
1 units (13/2 = 1)
31.6 units (103/2 = 31.6)
1000 units (1003/2 = 1000)
2828 units (2003/2 = 2828)
Notice that physical intensity grew from 1 to 200 units, but sensory intensity grew from 1 to 2828 units. Thus, even a small change in I produces a big change in S.
For a long time, Stevens' Law was held to be a
general psychophysical law that applied to all modalities, and
all qualities within a modality, under any circumstances. It
was assumed to describe the operating characteristic
(a term appropriated from engineering) of the sensory
These kinds of relations go far beyond the transduction of stimulus energies.
More recently, classical psychophysics, with its emphasis on thresholds, has been replaced by signal detection theory, which holds that detection is not merely a function of stimulus intensity, but also that the expectations and goals of the observer must be taken into account.
Still and all, psychophysics deserves to be
regarded as the beginning of the scientific approach to
consciousness, as it took mental states as its subject, and
had all the hallmarks of Enlightenment science:
The Visual Analogue Scale and Its Discontents
The intensity of a stimulus is often measured with a simple visual analogue scale (VAS) -- subjects are asked to mark a horizontal line to indicate how intense their experience is.
psychophysicists have criticized the standard
VAS on the ground that a rating of "7"
thing to one subject, and something entirely
different to another one (though, frankly,
Hilgard showed clearly that his open-ended
VAS for pain behaved quite well in
psychophysical terms). In order to
put subjects on a more common footing,
Linda Bartushuk, an expert on the
psychophysics and physiology of
introduced a General Labeled Magnitude
(gLMS), in which various
points on the scale are labeled with
verbal descriptors, as follows
(Bartoshuk et al., Currrent
Directions in Psychological
Science, 2005; illustration
from "Relief: Scientists Harness the
Power of Perception to Fade Out
Chronic Pain" by Rachel Adelson,
APS Observer, 28(9),
Bartoshuk was not, of
course, the first to label
points on the VAS.
But note that her labels do not
correspond precisely to the various
numbers" -- or, for that
matter, to our intuitions. A
rating of "Very Strong", for example,
is near the midpoint on the scale,
somewhere between 5 and 6.
They claim that
this form of VAS
has better psychometric properties than
the usual alternatives.
Classical psychophysics was largely focused on the problem of intensity, but even within the tradition of classical psychophysics, it was clear that there are other measurable attributes of sensation besides intensity. In principle, these different qualities of sensation are also subject to the psychophysical principle, in that each corresponds to some physical quality of stimulation. And, in principle, these different qualities of sensation also follow the psychophysical law, such as Stevens' Law.
But before we can get to psychophysical laws and principles, we need to know what these basic, primary, irreducible dimensions are. To a large degree, that task fell to the members of a school of psychologists known as Structuralism. who employed a method known as introspection to determine what these attributes were -- in other words, to identify what Edwin G. Boring, the great Harvard historian of psychology, called the physical dimensions of consciousness (Boring, 1953).
Introspection was introduced by Wilhelm Wundt, a German psychologist who founded the Structuralist school, and refined by E.B. Titchener, Wundt's American disciple, who founded the psychological laboratory at Cornell. Through the method of experimental introspection, stimulus conditions were carefully varied, and subjects were asked to describe their conscious experiences in terms of their constituent elements. The intended result was a sort of mental chemistry, in which sensory experiences, analogous to molecules, were analyzed into their constituent elements, analogous to atoms.
In their work, the Structuralists argued for a parallel between physics and psychology (they were, after all, psychophysicists!). These are known as the psychophysical parallels.
In classical Newtonian physics, there are three
(According to relativity theory, of course, space and time may constitute a single dimension of spacetime, but the structuralists were working before Einstein came onto the scene.)
Correspondingly, the Structuralists identified
three psychological dimensions analogous to the three
fundamental dimensions of physics:
In addition, the Structuralists identified two
dimensions that were unique to sensory experience:
While the classical psychophysicists, like Weber and Fechner, focused on sensory intensity, the structuralists focused on sensory qualities.
The sensory qualities, in turn, differed according to modality.
Some sense of what the structuralists were up to can be gleaned from a modern analog: wine-tasting, in which the taste of the liquid is broken down into its components. Consider the following example:
"...produced from estate grown grapes at RoxyAnn, the 2004 Syrah displays a dense, purple-black appearance with powerful, forward aromas of grilled plums and cracked black pepper accented with earth and tar notes. Dark and juicy, this wine is polished and smooth, with ripe integrated tannins layered over a chewy mouthful of concentrated blackberry, cherry, and white pepper. Lush perfumed spice and loads of ripe blueberry and blackberry flavors reveal lavender blossoms and a gentle touch of spicy oak on the long, lingering finish" (from an ad for RoxyAnn Winery in the Ashland (Or.) Glad Tidings, 09/15/06).
The example isn't perfect, not least because the structuralists would have sought to break down even these elements, such as "grilled plums" and "cracked black pepper" into their constituent elements as well. Moreover, describing a taste in terms of "blackberry" and "cherry" commits the stimulus error described below. But, still, you can get the idea.
The general idea of what the Structuralists were up to is given by the color circle, which shows how the four primary hues, varying in saturation, can be combined to yield a whole host of complex colors. Reversing the process, if you introspect on your experience of seeing a complex color, you can analyze it into basic elements of red, green, yellow, blue, and gray that can't be further reduced. And in this way you get the basic elements. These are qualia.
The examples that follow were constructed with PowerPoint presentation software, and they may show up a little differently on your screen, but you'll get the idea.
And so it goes. The structuralists would present their observers with complex sensory stimuli, and then ask them to analyze their experience into its basic constituent elements.
Introspection, a technique introduced by Wundt and refined by Titchener (who, in turn, was Boring's mentor in graduate school) required careful training of observers, so that they would avoid the "stimulus error" of describing the stimulus, and its meaning, rather than the qualities of the sensory experience to which the stimulus gave rise. For Wundt and Titchener, the stimulus error reflected a confusion between observation and inference -- and because introspection was construed as the empirical observation of conscious experience, inferences were to be avoided at all costs.
Of course, you can't report your introspections while you are making them, because the reporting will interfere with the introspecting. In his 1896 Outline of Psychology, Titchener made clear that all introspection is really retrospection (p. 36):
The rule of experimental introspection, in the sphere of sensation, runs as follows: Have yourself placed under such conditions that there is as little likelihood as possible of external interference with the test to be made. Attend to the stimulus, and, when it is removed, recall the sensation by an act of memory. Give a verbal account of the processes constituting your consciousness of the stimulus.
In whichever form it is employed, the introspective method demands the exercise of memory. Care must therefore be taken to work with memory at its best: the interval of time which elapses between experience and the account of experience must not be so short that memory has not time to recover the experience, or so long that the experience has become faded and blurred. In its experimental form, introspection demands further an exact use of language. The terms chosen to describe the experience must be definite, sharp, and concrete. The conscious process is like a fresco, painted n great sweeps of colour and with all sorts of intermediary and mediating lights and shades: words are little blocks of stone, to be used in the composition of a mosaic. If we are required to represent the fresco by a mosaic, we must see to it that our blocks be of small size and of every obtainable tint and hue. Otherwise, our representation will not come very near to the original.
Here's how Titchener captured the rules for introspection in his 1898 Primer of Psychology (pp. 33-35):
The rules for introspection are of two kinds: general and special. the latter refer to the regulation of stimulus, and differ in different investigations; the former refer to the frame of mind, and must be observed in all investigations alike.
Suppose, e.g., that you were trying to find out how small a difference you could distinguish in the smell of beeswax; that is, how much greater the surface of the stimulus must be made if the sensation of smell is to become noticeably stronger. It would be a special rule that you should work only on dry days; for beeswax smells much stronger in wet than in fine weather. Or if you were trying to discover how well you could call up the smell of beeswax in your mind, without having the wax under your nose, it would be a special rule that you should perform the experiment in a perfectly odourless room, so that the excitation set up from inside the brain should not be interfered with by foreign stimulations set up in the smell-cells of the nose. Again, if you were trying to distinguish all possible tints of blue, it would be a special rule that you should work always by the same illumination: always by dull daylight, or always by the same electric lights, etc. For a blue seen in sunlight is different from the same blue seen in dull daylight.
The general rules of experimental introspection are as follows:
- Be impartial. Do not form a preconceived idea of what you are going to find by the experiment; do not hope or expect to find this or that process. Take consciousness as it is.
- Be attentive. Do not speculate as to what you are doing or why you are doing it, as to the value or uselessness, during the experiment. Take the experiment seriously.
- Be comfortable. Do not begin to introspect till all the conditions are satisfactory: do not work if you feel nervous or irritated, if the chair is too high or the table too low for you, if you have a cold or a headache. Take the experiment pleasantly.
- Be perfectly fresh. Stop working the moment that you feel tired or jaded. Take the experiment vigorously.
And, above all, avoid the stimulus error, which Titchener described in his 1905 textbook of Psychology (pp. xxvi-xxvii):
We are constantly confusing sensations with their stimuli, with their objects, with their meanings. Or rather -- since the sensation of psychology has no object or meaning -- we are constantly confusing logical abstraction with psychological analysis; we abstract a certain aspect of an object or meaning, and then treat this aspect as if it were a simple mental process, an element in the mental representation of the object or meaning.... We do not say, in ordinary conversation, that this visual sensation is lighter than that, but that this pair of gloves or this kind of grey paper is lighter than this other. We do not say that this complex of cutaneous or organic sensations is more intensive than that, but that this box or package is heavier than this other. We do not even say, as a rule, that this tonal quality is lower than that, but rather that this instrument is flat and must be tuned up to this other. Always in what we say there is a reference to the objects, to the meaning of the conscious complex. It is not the grey, pressure, tone, that we are thinking of; but the grey of leather or paper, the pressure of the box, the pitch of the violin.... What is more natural than to read the character of the stimuli, of the objects, into the 'sensations' with which certain aspects of the stimulus or object are correlated...?
Of course, we could also say, as with James and Brentano, that if the goal of Structuralism is to produce a pure description of experience, then Structuralism misses the point. If mental states are about things other than themselves, then confining introspection to the mental state, and avoiding the object that it represents, isn't desirable -- or even possible. But let's set that point aside, and see where the Structuralist analysis of sensory qualia takes us. It turns out that it takes us to some pretty good places.
The Skinny on Introspection
Aside from the Titchener books mentioned in the text, here are three other good resources on experimental introspection:
Titchener argued that the elements of conscious experience could be grouped into two classes: sensations and feelings. He engaged in a dispute with Kulpe, another Structuralist, over the role of memory images, and whether there were ever imageless thoughts. Other structuralists argued that feelings were themselves sensations (you can see a shadow of this position in the James-Lange theory of emotion, which asserts that emotional feelings are actually the perception of bodily responses to a stimulus event). Still others asserted that images were sensations as well -- albeit excited centrally, though thought and memory, rather than peripherally, through stimulation and transduction. In any event, introspection focused almost exclusively on the qualities of sensation -- although, as we will see, Wundt later instigated a debate about the basic qualities of feeling.
In vision, there is:
In somesthesis, or the skin senses,
Boring (1933)` identified four fundamental qualities of:
The Structuralists also debated whether there were additional somesthetic qualities of roughness and wetness. The question, as always, was whether these were irreducible qualities of somesthetic experience, or whether the qualities of roughness and wetness were actually blends of elementary dimensions such as pressure and pain. More recently, the list of potential tactile qualia has been expanded to include soft vs. hard, tickle, and hot.
In taste, there appear to be four
basic sensations of:
In 1916, the Danish structuralist Henning, following up on earlier analyses by Hanig (1901), arranged these four gustatory qualities into a taste tetrahedron. In principle, according to Henning, the complex taste of any substance could be represented as a point in the space defined by the surfaces of the tetrahedron -- so much sour, so much salt, so much sweet, or whatever. Think of it the taste tetrahedron as a kind of "color circle" for taste.
As it happens, recent research (by Japanese investigators) has revealed a fifth basic taste, known as umami, which is a Japanese word meaning savory or deliciousness. Umami is a taste ingredient found in much Chinese and other Asian food, and is particularly present in monosodium glutamate (MSG) and soy sauce. It is also present in some meats and cheeses, though perhaps not in Denmark, which is why Henning didn't discover it.
In smell, there appear to be six fundamental sensory qualia:
Again, Henning arranged a smell
prism for representing the complex smell of any
substance as a point in the space defined by the surfaces of
Qualia refer to consciously perceptible qualities of sensation. But in the case of olfaction, there appear to be unconscious olfactory sensations -- namely pheromones (Karlson & Luscher, 1959). Pheromones are odorants, but they don't give rise to conscious olfactory sensations. Still, these odorants are known to affect the social behavior of a wide variety of animals, including humans. Martha McClintock's work on menstrual synchronization among women who live together in close quarters is a famous example.
Although taste (gustation) and smell (olfaction) are basic modalities of sensation, the flavor of any food is given by a combination of taste, smell, and touch. This is easily demonstrated by tasting a strawberry (without, of course, knowing that it is a strawberry -- otherwise you're liable to commit the stimulus error! -- while holding your nose. You will taste the sweet of the fruit, but you will not taste the flavor of strawberry until you breathe again. Similarly, irritants like carbonation, mint, and pepper contribute to flavor through the tactile modality. What pepper contributes to the taste of food is -- not to put too fine a point on it -- tactile pain.
In somesthesis, or the skin senses,
Boring (1933)` identified four fundamental qualities of:
Speaking of touch, Titchener (1920) was inspired by Henning to devise a tentative touch pyramid to map the relations among the various qualities of touch. As with smell and taste, any tactile sensation could be located somewhere on the surface of this pyramid. A somewhat different list of tactile qualities was advocated by E.g. Boring, Titchener's student, historian of psychology, and longtime professor at Harvard. We don't have to get distracted by the differences between teacher and student. Titchener's touch pyramid is offered here as just another example of the Structuralist approach to the identification of qualia.
Any such proposal, of course, is highly dependent on an accurate determination of just what these basic qualia are, and here the matter can get complicated. For example, just as there may be a fifth quality of umami in taste, so there may be a seventh quality of alkaline in smell. And we still don't know how many somesthetic qualities there are. Still, this level of uncertainty hasn't prevented the perfume industry from using the basic olfactory elements to synthesize new odors. similarly, the basic elements of taste are used by the food technology industry to synthesize new flavors (see "The New Spice Trade", The Economist, 06/22/02).
The Structuralists' introspection was not just an abstract exercise. Rather, it laid the basis for understanding the neural substrates of sensory qualia. Recall that in the early 19th century, Johannes Muller had proposed the Doctrine of Specific Nerve Energies -- basically, that each modality of sensation was associated with a separate neural system. Herman von Helmholtz, who had been Muller's student, proposed an extension of Muller's doctrine, the Doctrine of Specific Fiber Energies (1863, 1865) such that different neural systems underlay each quality of sensation with each modality (Helmholtz also came up with the idea of sensory modality itself).
So, at least in theory, each of different qualities of sensation is mediated by a different neural system which is dedicated to that purpose. As it happens, the neural bases of these different qualities has actually been worked out -- at least for some qualities, in some modalities.
The classic example, of course, is color, and
this is where Helmholtz initially propounded his
doctrine. Following Young (1802), Helmholtz proposed a trichromatic
theory of color vision, based on the fact, known from
physics, that all visible colors can be produced by a mixture
of just three basic colors:
Helmholtz then proposed that there were three
types of cones, each sensitive to a different part of the
visible electromagnetic spectrum:
Young and Helmholtz had the right idea, but
got the details wrong. The trichromatic theory was
eventually replaced by the opponent-process theory of
color vision initially proposed in the 19th century by
Ewald Hering, and revived in the 20th century by Leo Hurvich
and Dorothea Jameson. According to the theory, color
vision is based on three receptor systems, each maximally
sensitive to 3 classes of wavelengths, as in the trichromatic
theory. The output of these receptor systems, in turn,
becomes input to three opponent processes:
As stated by Hering, and by Hurvich and Jameson, opponent-process theory was a psychological theory, inferred from the data of human performance. More recently, however, investigators such as the late Russell DeValois at UCB, have identified the physiological substrates of the opponent processes.
example, somewhat more speculative at this time, concerns the
relation between taste qualities and taste buds on the
tongue. Histological analyses of the tongue have
identified three different papillae, or taste buds:
In principle, the output of these papillae could
be integrated to yield all the different tastes we can
experience, somewhat along the lines of the opponent-process
theory. Note, however two flies in the ointment:
For a long time, it was thought that receptors specific to the four "basic tastes" -- sweet, sour, salty, and bitter -- were located in discrete areas of the tongue. The sensation of sour is most acute at the sides, least at the tip and base. The sensations of sweet and salty are most acute at the tip, least at the sides, and base. The sensation of bitter is most acute at the base, and on the throat and the palate. This suggests that there may be four different receptors, each sensitive to a different basic taste, and each uniquely distributed on the tongue -- after the manner of the "tongue map" proposed by E.G. Boring in 1942.
Unfortunately, histological studies have failed so far to reveal differences in the taste buds at various sites on the tongue, so the precise organization of taste remains something of a mystery. While it's true, to some extent, that different areas of the tongue are differentially sensitive to the various basic tastes, the situation is more complicated than this -- not least by the discovery of umami, that fifth basic taste, which made Boring's "taste map" more than a little quaint, much like maps of the world printed before 1492. More critically, it appears that clusters of cells, each sensitive to different chemical molecules, are dispersed widely across the tongue. Much in the manner of color vision, these clusters may be maximally sensitive to particular molecules, they are also responsive to others. And much in the manner of color vision, the output from these receptors seem to be integrated at high levels of the nervous system to yield the sensation of taste.
But before we can go looking for the neural substrates of different sensory qualia, we have to know what those qualia are -- the psychology always precedes the neurology. Which begs the question: how do we know a basic sensory quality when we see (or hear, or taste, or smell, or touch) it?
Cutting (2008) has used the case of color or hue as models for
identifying basic qualities in other sensory modalities.
In his view, a basic quality has the following properties:
It would be an interesting
exercise to apply Cutting's criteria to proposals for basic
qualities in other domains (hint, hint).
Consider, for example, some
recent cross-cultural work on odor-identification by Asifa
Majid and her colleagues at the University of Nijmegen in
The Netherlands, who have studied the odor lexicons of
various Aslian languages spoken by indigenous tribes on the
Malay Peninsula (e.g., Majid & Kruspe, Current
The point of all this is to suggest that English-speakers
may not be the best subjects to use for studies of sensory
qualia outside of vision and audition. For
hunter-gatherers like the Jahai and the Semaq Beri, odors
are codable in abstract, monolexemic terms, as Cutting
suggest. But for members of developed Western
societies, we've somehow lost that code, and have to make do
with the source-based labels that are left to us -- and
these may not match up to the actual qualities of sensory
Much as Cutting
(2008) derived some ideas from the basic colors,
Ramachandran and his colleagues have been inspired by
synesthesia (e.g., Ramachandran & Hubbard, 2001).
They call these principles New Laws of Qualia:
structuralists focused on sensory experience, in
principle the program of introspection, identifying the basic
dimensions of conscious experience, could be applied to other
elements as well. In fact, Wundt himself proposed that
there were three dimensions of emotional feeling:
Titchener countered with a different proposal, that pleasantness-unpleasantness was the sole dimension of feeling.
We can see in this system rough analogies
to Wundt's three primary psychological dimensions of
extensity, intensity, and protensity -- and thus to the
fundamental dimensions of space, mass, and time in classical
Newtonian physics. Wundt was nothing if not
consistent. Note, however, that while the sensory
attributes are unidimensional (at least, intensity and
protensity vary from zero to infinity), Wundt's dimensions of
feeling are bipolar. For Wundt, pleasantness and
unpleasantness were opposite poles of a single
dimension. A "zero" degree of pleasantness, like a zero"
degree of unpleasantness, lies at the midpoint of the
dimension. Similarly, calm is not just the
absence of excitement, and relaxation is not
just the absence of strain: they are somehow
Wundt's assertions in this regard were debated by Titchener, who argued that pleasantness-unpleasantness was the sole dimension of feeling. Titchener agreed, however, that pleasantness and unpleasantness are negatively correlated, comprising opposite poles of a single continuum..
This is the conventional way of thinking
about the structure of affect: you can't feel bad while
feeling good. For example, Russell summarized research
on the affect lexicon -- the language of emotion -- in his affect
circumplex. Technically, a circumplex is a
circular arrangement of variables, each represented as a
vector in space, where the angular distance between vectors
represents the correlation between the variables they
represent. For example, a small angular distance means
that the two variables are highly positively correlated; a
large angular distance means that the two variables are highly
negatively correlated; and a right angle means that the
two variables are not correlated at all. In Russell's
circumplex, there are two dimensions: positive-negative, and
strong-weak. In the circumplex, positive and negative
affect are opposite ends of a single dimension:
Russell's affect circumplex is essentially Wundt's solution to the structure of affect, imported from the late 19th century into the late 20th.
Later, Watson and
Tellegen argued for an alternative circumplex, in which
positive and negative affect are independent of each
other. In the Watson-Tellegen circumplex, there are two
big dimensions, one of strong-to-weak positive affect, and one
of strong-to-weak negative affect. In principle, it is
possible to be in a state characterized by high degrees of
both positive and negative affect -- as in states of high
arousal, astonishment, or surprise.
Just as the 19th-century structuralists got into endless debates about the structure of sensory experience, so 20th- (and 21st-) century emotion researchers have gotten into a seemingly endless debate about the structure of affect.
For example, affect circumplexes are generated by a statistical technique known as factor analysis, which is based on the correlation coefficient. For technical reasons that need not detain us here, there is a certain amount of subjectivity in the placement of principal factors. Note, for example, that in the Watson & Tellegen affect circumplex there is, in fact, a bipolar dimension of pleasantness-unpleasantness, just as Wundt and Russell maintained. It's just that Watson & Tellegen believe that the independent factors of positive and negative affect are stronger, and dominate the statistical solution. A further problem, unfortunately, is that correlation coefficients are not perfectly reliable, and seemingly minor differences from one study to another can lead to major discrepancies in the placement of factor axes. It does not help, either, to note that Russell and Watson & Tellegen factor-analyzed somewhat different sets of emotion terms.
For these and other reasons, both substantive and non-substantive, an enormously vigorous dispute over the correct structure of affect continues to this day.
For example, Donald Green, a political
scientist interested in political attitude polling, and Peter
Salovey, a psychologist interested in emotion, have argued
that the Wundtian structure is the correct one after
all. In one study, Green, Goldman, & Salovey (1993)
factor-analyzed affect ratings, using a multi-trait multi-method
methodology to reduce measurement error, and obtained solid
evidence that happiness and sadness were opposites, rather
than independent constructs.
Tellegen and Watson
(1999a, 199b), for their part, offered a new hierarchical
factor analysis that provides yet a third perspective on the
structure of affect. Analyzing the relations among 27
3-item scales measuring 9 content categories of affect,
Tellegen and Watson found two strong "secondary" factors,
independent of each other, tapping positive and negative
activation -- essentially confirming their earlier
results. However, a further analysis yielded a single
"tertiary" factor of happiness-unhappiness -- essentially, the
bipolar structure favored by Wundt, Russell, and Green and
Salovey. So, the structure of affect appears to depend
on the level of analysis.
Determining the basic qualities of affective
experience seems a necessary first step in determining the
neural substrates of those mental states.
But before we go looking for brain systems, we're going to have to have some idea what we're looking for. If Ekman is wrong, and there are more than a half-dozen basic emotions, then we'll never find all the brain systems that are involved in our emotional lives.
Consider, for example, the case of pain. Studies of clinical patients (and laboratory subjects) experiencing pain, using such instruments as the McGill Pain Scale (Melzack & Torgerson, 1971; Melzack, 1975) reveal that the pain experience comes in several different kinds. Each of these kinds of pain appear to be mediated by a different neural system.
First, there are two components to the pain
Similar studies have identified two different
types of sensory pain:
Apparently, however, both fast and slow pain signals end up in the somatosensory cortex.
These proposals are somewhat controversial; and, as in the Young-Helmholtz theory of color vision, they may be wrong. But we'll never find the neural substrates of sensory qualia unless we know that the qualia are first. The psychology comes first.
Why did the structuralist program fail? Partly, it was cut short by the onslaught of behaviorism. Partly, it was eaten from within by debates over things like imageless thought. But more important, the structuralists' fundamental assumption -- that conscious experience could be analyzed into its sensory elements -- was undercut by the Gestalt psychologists who showed that, in perception, "the whole is greater than the sum of its parts". But setting aside these debates, the program was sidetracked by its focus on qualia rather than intentionality. By rigorously training introspective observers to avoid the stimulus error, the structuralists neglected the fact that mental states are not pure and abstract experiences; they are always "about" something.
This page last modified