The Windows of Perception
A driving conceit of modern psychology is that the brain somehow creates consciousness. Of course, no one is quite sure how this is possible. However well one’s theory of consciousness explains the still developing data, there still remains the nagging question as to why that thing or process could result in true mental states. For example, given an algorithmic account, in which the brain is conceived as the hardware on which neural-based software runs, the question is how a particular algorithmic step or series of steps could result in the smell of a rainy day, the exhilaration associated with musical pleasure, or for that matter, the simple act of seeing blue. The thing being explained and the explanation just seem too conceptually distant for the explanation to get any traction. This has been labeled the hard problem of consciousness (Chalmers, 1996).
One way out of this conundrum is to relieve matter/process of a power that it seems unlikely to possess. Certainly, what the brain does affects the nature of consciousness, but that does not imply that it generates consciousness, any more than a prism creates light, or a circuit makes electrons. In these cases, the material process serves solely to transform a pre-existing substance. Likewise, this paper will make the following about the interaction between consciousness and the brain:
Pure consciousness exists outside of space-time, and is neither created by the brain, or any other material process, although such processes may serve to differentiate consciousness into various subsets of its full form.
In other words, the brain is a system that works as a filter, which, on the basis of perceptual input and internally-generated thoughts, selectively allows aspects of pure awareness to be experienced.
Thus, the central assumption entails a kind of essentialism, in that consciousness is neither created nor destroyed but is part of the universe, ab initio, and also a kind of dualism in that consciousness is not directly identified with the mechanisms which act on it. However, unlike earlier theories of this type, the goal here is not merely to argue that the assumption is a philosophical possibility, but to develop the assumption along scientific lines.
In particular, the central question that this paper attempts to answer is what aspect or aspects of neural functioning control the degree to which pure consciousness filters into a particular awareness. It will first be argued that the brain acts to separate pure consciousness, which by hypothesis is a non-dimensional entity, into information that resides in an ordered n-dimensional space. The carrier of this information is the causal interaction between the elements of neural processing. However, given the appropriate stimuli, this ordering may break down, and a glimpse of original undivided consciousness may be had. Two sources of data will then be considered to confirm these ideas, neural models of aesthetic perception, and the neuroimaging data on meditative and other similar states. Taken together, these data suggest that the brain, rather than being the originator of the mental, in fact limits access to consciousness in its undiluted form.
The Dimensionality of Experience
If pure consciousness “resides” apart from space-time, and instantiated consciousness or consciousness of a particular item exists in a particular dimension, the question naturally arises as to how dimensionality arises from the lack thereof. This is a difficult problem, and indeed, a complete solution of which would throw considerable light on the entire problem of consciousness. Nevertheless, a reasonable approximation to an answer may be attempted with reference to the ordering of visual experience. In addition to providing a clear example of ordering in phenomenal space, vision makes an excellent example for this issue because it is the single most studied modality from the point of view of both psychology and physiology.
Three possible means by which visual experience attains its spatial character are now considered:
- Phixels (phenomenal pixels) are ordered by attached or implicit indexing information.
From a programming point of view, this is perhaps the most natural method. If one were to program a two-dimensional field of colors, for example, then it is probable that one would use a two-dimensional matrix of pixel values. The position in space a pixel would then be its associated indices, and distance relations between this pixel and others would be determined by the difference between the indices. This method, while ideal for the programmer, has one serious drawback as an explanation of the ordering of visual experience. This is that it is entirely passive, but consciousness is an active process. By this is meant that unless a particular pixel is being operated on, then the information regarding ordering is not being accessed. Thus, it cannot contribute to the overall parallel ordering at any given instant, which is characteristic of the visual experiential field.1
- The so-called “where” system in the parietal cortex is responsible for phixel positioning.
It has been suggested that the ventral projection from the primary visual cortex to area IT is responsible for the recognition of the visual stimulus, and that dorsal projections from the visual cortex to the parietal cortex subserve positional information (Ungerleider and Mishkin, 1982). However, it appears that parietal calculations are invoked in order to accomplish specific motor actions (Milner and Goodale, 1995). Thus, these could not be the locus of full positional information because only limited motor actions are contemplated at any given time. Furthermore, it is known that actions such as grasping are immune to the perceived position of objects. For example, if one is presented with a visual illusion, one grasps at the actual elements rather than the perceived elements, indicating that consciousness is not affected directly by parietal calculations (Post and Welch, 1996).
- Causal relations determine phixel ordering.
One remaining possibility that is consistent with the idea that the contents of consciousness are constructed in the here and now in a parallel fashion is the notion that causality determines visual ordering. Causality is the basis of functionalism, a popular theory of mind that claims that qualia are a function of the causal transformations from sensory inputs to motor outputs within the brain, although here the claim is much more specific. This claim is that the phenomenal distance between two visual qualia is inversely proportional to the causal currents between the neurons subserving these experiences.
What evidence is there for this? The primary reason to believe this with respect to visual perception is that a) visual cortex is arranged retinotopically, and b) that lateral connectivity in the visual cortex is typically a decreasing function of distance. It follows immediately from these two propositions that neurons subserving proximal visual regions are also the ones with the greatest mutual causal influence. The remaining step is to invoke the intimate connection between causality and consciousness to support the claim that it is the causal relations themselves that establish the visual field.
Generalizing this conception from visual ordering to ordering in general leads to
Causal interactions in the brain induce an ordering in one or more dimensions
The second related premise can now also be stated.
The collapse of this ordering leads to aesthetic pleasure, or in the extreme case,the apperception of pure consciousness in its undiluted state.
The justification for this premise is the aforementioned claim that consciousness it its original form is a dimensionless quantity. In ordinary awareness, causality serves to separate out the elements of perception and place them in an ordered sequence. However, in order to do so, causality must suggest an unambiguous sequence. To take a simple example, suppose A influences B and B influences C, then we have the ordering A – B – C. However, if C also influences A then any ordering imposed would be arbitrary, as each element is connected to the other two. This collapse of ordering is presumed to lead to aesthetic pleasure, which by this conception is a glimpse at the coattails of pure unadulterated consciousness.
The prior derivation of Premises 1 and 2 is admittedly sketchy, but a complete treatment would require many pages, and is thus withheld. Instead, it is hoped that the prior discussion suggests how a fuller justification might proceed, and that the discussion in the following section will serve to bolster and further illuminate these notions.
Applications of the model
Here we consider three applications of the model, humor, visual art, and music. In each, it will be shown that aesthetic pleasure is intimately tied to the breakdown in the normal ordering associated within either a given modality as in the latter two applications, or within conceptual space itself, as in the following application.
Humor may seem an unlikely place to begin to demonstrate the applicability of a theory of consciousness, but it turns out that it has many desirable characteristics in this regard. Humor invariably consists of the priming of a given frame or set of expectations, and then the rapid shift to an alternative set. Furthermore, in order to work, this alternative must evoke strong cognitive support. In general terms, then, we have one conceptual framework supported by expectation, and an alternative framework supported by the context and other facilitating details. If conceptual space is an ordered space in which related concepts are placed near each other, and with more distant ones farther apart, then humor works by rapidly fusing a temporary connection between these ordinarily distant ideas.
As we are on the topic, we may as well use a consciousness joke as an exemplary instance:2
First behaviorist to second behaviorist after making love: “It was great for you, how was it for me?”
This joke is a play on the nature of qualitative experience, which is ordinarily considered to be private and only indirectly accessible to third parties. However, for a behaviorist, such private states are nonexistent. Hence, the absurd but plausible attempt by the first behaviorist to gain access to his internal states by querying someone else.
Figure 1. Two ordinarily distinct concepts, “qualia” and “public” are temporarily connected by the joke context.
In my neurally-inspired model of humor (Katz, 1993), this joke would be rendered as in Figure 1. Within a hypothetical conceptual space, the concept of qualia (or qualitative experiences) is closely tied to the notion of privacy, with the opposing concept of public experience distantly represented. However, the joke temporarily makes it plausible that qualia should be treated as public data. In the model, the context of behaviorism is connected to both of these ordinarily opposing notions, forming a virtual temporary link between them (dotted line). This link breaks down the normal ordering whereby the notions of public and private cannot jointly be entertained.
An additional reason that the above joke works is that it contains tendentious elements, in this case aggressive ones (poking fun at the behaviorist for thinking this way), and the obvious sexual element. These elements are not directly implicated in the breakdown of ordering but may be seen to facilitate this breakdown as follows. Both sexuality and aggression (as long as it is directed toward the appropriate target) facilitate the salience of stimuli of which they are elements. This may be plausibly be modeled as a kind of searchlight that serves to provide extra activation to these elements. This additional activation may serve to overcome mutually inhibitory links between opposing concepts such as private and public (not shown in the diagram), and foster a stronger temporary connection between distant conceptual spaces.
In summary, humor fuses together elements that are normally distant in conceptual space, upsetting the natural ordering of this space. This process may be amplified by tendentious components, which are almost always present in humor. In the next section, it will be suggested that visual art works in a similar albeit less radical manner.
Visual perception provides perhaps the most direct confirmation of the current proposal because visual perceptual space is naturally ordered. If it were not, colors and forms would be superimposed upon each other in an unintelligible muddle. Instead, there each perceptual pixel can be cleanly located in a grid of such elements. What is not always appreciated, however, is that the construction of this perceptually space can be influenced by the elements of this space. In a remarkable series of experiments, Coren and Girgus (1980) showed that the perceived distance between visual elements is a function of the Gestalt formed by those elements. Consider the dots in Figure 2 forming a square, for example. The distance between any two dots within this form is generally perceived as between 4 and 10% closer to each other an objectively identical distance not within the form.
Figure 2. The internal distance is usually perceived as being less than the external distance despite the fact that they are equivalent.
Thus, perceptual space is malleable, and is affected by among other things, the goodness of form, or Gestalt of the elements in that space. We can say, therefore, that a work of visual art that works has a more compact feel to it that one that does not. This conforms to the principle outlined above, namely that the aesthetic response is a function of the collapse of a normally well-ordered space. Consider, for example, the two objects in Figure 3. These are taken from the visual aesthetic sensitivity test (VAST), designed by the psychologist H.J. Eysenck in collaboration with the German painter Götz et al. (1979). The object on the right is the aesthetically “correct” one, with the one on the left violating a number of basic principles. In particular, the indentations to the left and right are not congruent, and in addition the edge on the top right is much sharper than the others in the figure. In contrast, the figure on the right shows greater congruency between both the sides and the tops and the bottom. The dotted lines indicate the hypothetical causal interactions within the mind of the perceiver over and above the normal spatial ordering in any 2D plane. It these interactions which are suggested to render the figure more perceptually compact than it would be otherwise, and therefore contribute to the collapse of visual ordering. In general, it is suggested that visual art works by pulling together in the mind of the perceiver normally distant spatial elements via Gestalt principles such as similarity.
Figure 3. The preferred figure in B forms a better Gestalt with greater similarity between the horizontal and vertical sides.
In prior studies, I have shown that musical passages that people find pleasing lead to an increase in synchrony of firing over the neural elements that subserve these passages (Katz, 2004). Here I extend this model to show a partial collapse in ordering during the processing of pleasing music. Figure 4 shows two dominant to tonic transitions, or perfect cadences. However, the one on the left follows the principles of voice leading, in which notes make minimal jumps, while the one on the right does not. The dominant note g in the former is constant throughout; in the latter it jumps an octave. Listeners generally prefer the former transition.
Figure 4. Two possible dominant to tonic chord transitions. The one in the first measure follows principles of voice leading and is therefore preferred to the one in the second measure.
A simplified neural model to account for this is shown in Figure 5. Here units in the note layer connect to the chords of which they are members in chord layer, with the strongest connection (thicker line) extending from the tonic of the chord. The bottom of Figure 5 shows what happens when there is good voice leading. Here the shared g keeps the dominant, V chord unit and its constituent notes partially active while the tonic, I chord unit is triggered. This establishes an indirect temporary connection between the notes in the two chords, for example, the e and the b (dotted line). This, along with the connections between notes within a chord (not shown), leads to a breakdown in the normal linear ordering between the notes, and is hypothesized to lead to musical pleasure.
Figure 5. (Top) A model of chord recognition. (Bottom) Activation in the model for a transition following good voice leading after processing the tonic (I) chord. The shared note g between the chords establishes a virtual postive connection (dotted line) from the e to the b, among other things.
It is not difficult to extend the model to other aspects of music. Consider that the melodies in successive phrases of songs usually share a great deal of similarity. For example, the first two phrases of the song “Amazing Grace” have the same rhythm, and the first phrase starts with the note the second phrase ends with. These similarities disrupt the ordinary linear flow of time by binding together these temporally successive events.
Related experimental findings
Let us now address the relation between recent results on the state of the brain in the meditative state as measured by EEG and the proposed theory. Two main findings characterize these studies a) meditation leads to greater coherence between brain regions (Travis and Wallace, 1999), and b) meditation leads to an increase in gamma activity (Lehmann et. al., 2001). The former result is a direct consequence of the model, if it is assumed that EEG coherence is a measure of cortical interaction. The latter also follows from the model, if gamma activity serves to bind percepts, as has been suggested (Gray, 1999).
While it is still too early to form definitive conclusions from these preliminary results, both strongly point to the idea that causal interaction between brain modules is correlated with meditative state. If this state is not merely one in which the mind is calm, but also an active one in which the mind is better able to perceive its original, undifferentiated form, then increased causal interaction could be the explanation the apperception of pure consciousness, as Premise 2 suggests. It is also instructive to look at this from the opposite point of view, that of less than average coherence. For example, Tononi and Edelman (2001) have suggested that schizophrenia is the result of an ordinarily coherent consciousness breaking down into weakly communicating sub-modules. Further study is clearly needed to understand both pathological and exceptional brain states and their relation to the mental coherence.
In summary, brain processes are not assumed to generate consciousness, but to regulate it, both in the particular form that it takes at any given moment, and the amount of pure consciousness that is experienced. This regulation is governed by the nature of causal interaction between processing modules in the brain. In quotidian awareness, this interaction is localized, and perceptual inputs are ordered relative to each other. For example, visual perception is characterized by each percept occupying a unique position relative to other visual percepts. Were this not so, then survival would not be possible, and thus there is strong incentive to view ordinary perception as the default and desirable state of things. However, the differentiating aspect of the brain is antithetical to the experience of the full nature of consciousness. Cultures have therefore also evolved special practices such as music, meditation, and religious practices to temporarily suspend this differentiation, and to permit the experience of the original state of consciousness, which is unbounded in space-time.
Beardsley, M. C. (1966) Aesthetics from classical Greece to the present. New York: Macmillan.
Chalmers, D.J. (1996). The Conscious Mind. New York: Oxford University Press.
Coren, S. & Girgus, J.S. (1980). Principles of perceptual organization: The Gestalt illusions. Journal of Experimental Psychology: Human Perception and Performance, 6, 404-412.
Engel, A.K & Singer, W. (2001) Temporal binding and the neural correlates of sensory awareness. Trends in Cognitive Sciences, 5, 16-25.
Götz, K.O., Borisy, A.R., Lynn, R., & Eysenck, H.J (1979). A new visual aesthetic sensitivity test. I. Construction and Psychometric Properties. Perceptual and Motor Skills 49, 795.
Gray, C. (1999) The temporal correlation hypothesis of visual feature integration: Still alive and well. Neuron, 24, 31-47.
Lehmann, D., Faber, P.L., Achermann, P., Jeanmonod, D., Gianotti, L.R., Pizzagalli, D. (2001) Brain sources of EEG gamma frequency during volitionally meditation-induced altered states of consciousness and experiences of the self. Psychiatry Research, 108:2, 111-121.
Katz, B. (1993) A neural resolution of the incongruity and incongruity-resolution theories of humour. Connection Science, 5, 59-75.
Katz, B. (2001) What makes a polygon pleasing? Empirical Studies of the Arts 20:1, 1-19.
Katz, B. (2004) A measure of musical preference. Journal of Consciousness Studies, 11:3, 28-57.
Milner, A.D., & Goodale, M.A. (1995) The visual brain in action. Oxford: Oxford University Press.
Post, R.B., & Welch, R.B. (1996). Is there dissociation of perceptual and motor responses to figural illusions? Perception 25, 569-581.
Tononi G., Edelman G.M. (2000). Schizophrenia and the mechanisms of conscious integration, Brain Research Reviews, 31:391-400, 2000.
Travis, F. & Wallace, R.K. (1999) Autonomic and EEG patterns during eyes-closed rest and transcendental meditation practice: the basis for a neural model of TM practice. Consciousness and Cognition 8:3, 302-318.
Ungerleider, L.G. & Mishkin, M. (1982). Two cortical visual systems. In D.G Ingle, M.A. Goodale, & R.J.Q. Mansfield (Eds.) Analysis of visual behavior (pp. 549-586). Cambridge: MIT Press.
2 There are very few consciousness jokes, but here is another one: Why did the zombie chicken cross the rode? For exactly the same reason the sentient chicken did.