Simulating another's point of view is done by assuming different perceptions of the world. But even though this says "of the world", the ability to take into account perceptions "of the world" in other increasingly "distant" situations requires (unseen) mastery of the perception/simulation mechanism.
0: Direct perceptions of the world
1: Perceptions of what "just happened" in the world in the current situation
2: Perceptions of what happens in the world in another situation (but previous perceptions don't change how future perceptions are interpreted)
3: Perceptions of what happens in the world in another situation, interpreted based on having been exposed to different perceptions
4. Perception that running different simulation is all there is to the experience of existence. There is no core or container.

Orders of Perception

by Jeff Thompson


What is needed for a mind to fully perceive the structure of perception? Perception is a process which has 5 components, numbered 0 through 4. The mind needs to develop through stages to be able to perceive these components. Many theorists studying the mind, such as artificial intelligence researchers or cognitive scientists, are only able to perceive components 0 through 3 and are at a stage of development where the mind rejects perceptions of component 4. In part, this paper argues that if such a theorist wants to understand the mind, it is necessary to overcome the rejection of perceptions of component 4. Once all the components can be perceived, the processes of mind can be studied and worked with fully.

Perception As Prediction and Test

"Perception" here is used in the most general sense as anything that registers in the mind. In general, the mind perceives something if it can distinguish whether it is there or not. Thus "perception"  includes the simple case of seeing if a toy block is there or not, to distinguishing that the number 245 is odd, to the more abstract case of perceiving that this paper expresses an instance of a theory of knowledge.

An important point is that, in perception, the mind is experiencing a process that it itself is generating, and a theory of perception has to be about that process. If a person is looking at a ball, the person doesn't perceive the "actual" ball, even though that's what it seems like. The actual ball is a buzz of subatomic particles somewhere out there in the continuum of matter/energy, some of which interact with light which makes a pattern on the person's retina in the eye. So, the person doesn't perceive the actual ball. The person doesn't even perceive the pattern of light on the retina. Rather, the person's mind actively generates an image which is kept up to date by interaction with the retina, and it is only the image "in the mind" which is ever experienced. Loosely speaking, the mind is always hallucinating (but hopefully the hallucinations are coordinated with the reality of the objective world). There are many examples which could reinforce this - such as the eye's blind spot, visual illusions, and the fact that we "see" in dreams - but we aren't going to go into them here. Instead we take it as given that in perception a mind is experiencing a process which it itself is generating.

The mind perceives something if it can distinguish if it is there or not, so how does it make this distinction? By making predictions and testing. In perceiving a shape, for example a square, detecting an edge triggers a hypothesis that it has that shape. "Having that shape" is a set of predictions such as "edges will go in a straight line," "following any edge will come to a corner," "following all the edges and corners will connect around a perimeter where there are two pairs of corners," etc. The actions in each of these predictions causes a test which can fail. The eye actually moves around the shape performing the tests. If none of the tests fail, the hypothesis prevails and the mind has the experience of perceiving the shape. If any of the tests fail, another candidate hypothesis that also might be triggered by detecting an edge (such as a triangle shape) is tested and if it doesn't fail then that is the perception which prevails.

Note the order in which this occurs. A common misunderstanding is that perception begins with finding corners and edges, "processing" these, and concluding "if it has these corners and edges then it must be a square shape." But the "if then" of this process is actually the other way around. The detection of some edges and corners may trigger a hypothesis. But the point is that perception begins with a hypothesis like "if it is a square shape then it should have these edges and corners" and from that predictions are generated to test the hypothesis.

Besides making predictions which have tests that don't fail, there is another way for a hypothesis to prevail: it can prevent a threatening test from being run in the first place. This is called "blocking." Since there is limited capacity for running tests, blocking can be accomplished by forcing "friendly" tests to be run repeatedly. In this way, an unworthy hypothesis can be made to prevail (so long as there is not a higher order perception that this blocking is being used, which we'll discuss below.)

Structure of Perception

So if perception has this "prediction and test" structure, what are the components which must be present for this process to work? There is a causal chain in this process among a finite set of components. We begin with the final component of the chain and proceed back "up" the causal chain, asking at each step what is needed to bring about that component.
  1. Measurement: Refutable assertions about what is present. Measurement is the "atomic" operation in perception on which the other components are layered. For example, in simple visual perception, a measurement is "it is brighter in the middle than on the outside", "there is a vertical edge here crossing a horizontal edge." Or in auditory perception "it gets quieter after the beep". Each of these is an assertion which can be refuted.

    But if an assertion like "there is a vertical edge here crossing a horizontal edge" is present, how did it become present? Through some action, such as moving the eye, which brought the assertion under test. Likewise the assertion "it feels smooth" is present because of the action of running the finger along a table top. So if a measurement "comes from" an action, then is "action" the next component? Not quite, because if the mind just perceives an action like "the eye is moving" or "the finger is touching the surface" then that would be another atomic assertion. The action is only important because it brings an assertion under test such that if the assertion is refuted it means something. So, what does the mind perceive if it perceives the connection between action and measurement?

  2. Prediction: Action predicts measurement. The mind perceives a prediction through the success of actions which keep measurements present which are not refuted. The environment is constantly bringing about new conditions so that a measurement which was previously unrefuted is now refuted, and so action is necessary (e.g., shifting the eye or moving a hand) to bring about a  new measurement which is not refuted. There could be many actions, but only one can be performed. The mind not only perceives that this action is performed as opposed to another, but that the action completes upon the presence of the measurement - i.e., that the action was "meant for" the result (predicts it).

    So, if a new set of conditions brings about a prediction that an action will result in a measurement, is "set of conditions" the next component? Not quite. The mind may perceive the conditions (e.g., the table is tilted and the ball is rolling) and may perceive the action and predicted result (e.g. put the hand in front of the ball and the ball is in the hand) but not necessarily perceive how the prediction applies to the conditions. So, what does the mind perceive if it perceives the connection between conditions and the predictions that apply to it?

  3. Theory: Set of conditions implies certain predictions. Whether or not a mind is able to perceive a theory, it is always the case that a new prediction (and its actions) are brought about by a new set of conditions (perhaps instinctually or through simple associations between the conditions and the prediction). But if a mind can perceive a theory, then it can distinguish whether the theory is there or not, i.e., whether the actions are being performed as a result of a prediction that under these set of conditions, the actions would bring the results. Being able to perceive a theory also brings the ability to run "what if" scenarios, e.g., "if the conditions were different, what results would these actions bring?" If we consider that the set of all possible theories are divided into groups of conceptual categories, then, e.g., perceiving Rover as a dog means perceiving that Rover is an instance of the conceptual category "dog", which means that all the predictions in this category apply to Rover.

    A mind can perceive a rich world of theories with subtle shades of conceptual categories, and make fine distinctions about which theory may apply to a set of conditions. But why does the mind have these theories and not some others? A theory is based on a set of assumptions, so is "assumptions" the next component? Not quite. The mind may perceive a set of assumptions, but just as more theories or concepts. So, what does the mind perceive if it perceives the connection between assumptions and a theory?

    A perceiver does not react to the object in the world but to the activated model in the perceiver's mind. If a mind is not able to form or activate this theory, the perceiver cannot not react. [Which level is able to perceive that causality must pass through the activation of a theory?]

  4. Dependency: Assumptions yield a theory. "Assumptions" is a general term for the basis of a theory and includes logical premises as well as historical contingency. E.g., a mind may have a set of beliefs based on the mere "assumption" that it is brought up in a culture where only those beliefs are found. "Dependency" is a general term and includes logical necessity as well as mere blind trial and error. Whether or not a mind is able to perceive a dependency, it is always the case that theories have some specific path by which they obtained their specific form (even if they are generated by random mutation and are refined by natural selection). But if a mind can perceive a dependency, then it can distinguish whether the dependency is there or not, i.e., whether the present theory derives from the assumptions. Being able to perceive a dependency also brings the ability to take on hypotheticals, e.g., "what different theories or world view would a mind operate under if these other assumptions were true?" Instead of locating itself among a set of concepts, a mind which perceives dependencies can locate itself in the space of logical possibilities, from which particular concepts are easily derived if necessary. (E.g., "if you are a physical reductionist, then you necessarily hold to the following theories....")

    A perception depends on the path that developed it, but why do perceptions develop at all? What is the driving mechanism? In order for a perception to prevail, it must run all its predictions which distinguish it from alternative perceptions. Loosely speaking, it must permit each test in order to be sure it's "really the one." But the environment is always shifting and no finite perception can fully characterize it, so there is always a chance a test will fail. As discussed above, another way to make the perception prevail in this case is to block threatening tests. If this tactic worked, it would prevent development. But shutting down testing conflicts with the process the perception uses to fully distinguish itself, so it must again permit testing. This is the driving dynamic. So is this "blocking and permitting" the next component? Not quite. A mind could perceive the process of blocking and permitting without seeing how it forms the current perception. So, what does the mind perceive if it perceives the connection between blocking/permitting and dependency?

  5. Enforcement: Perceptions develop in the dynamic between blocking and permitting tests. As with all the components, enforcement operates in every perception, even simple ones. But this is called "enforcement" because, if a mind is able to perceive enforcement, then it can distinguish whether enforcement is there or not, i.e., whether a perception prevails due to blocking the tests which would arise as it is asserted. Perceiving this process, a mind may "override" and permit a test that would otherwise be blocked. An important example is the perception of continuously existing over time (which is needed to make identification with one's own personality feel real). Of course, this appearance is enforced by a perception which only momentarily activates. But perceiving that the perception only momentarily activates has implications for the felt reality of the personality and are usually blocked. Only a mind that can perceive enforcement can perceive that the present perception is blocking tests and choose to permit them.

    Is there a further component beyond enforcement that causes it? Enforcement comes from the necessary fact that a finite perception will fail to fully characterize its environment - but also that a perception which doesn't "try" to do so will not prevail. (The perception does not intentionally "try" any more than the species in "survival of the fittest" tries to survive. It is that the species which prevail appear to have tried to do so.)  This is supported by general physical mechanisms which provide the environment in which successful perceptions prevail, but since these mechanisms are general, they are not specific to a particular perception, and so cannot be considered a "component" of that perception. Therefore enforcement is the final component of a particular perception.

Orders of Perception

The main thesis: All these components 0 through 4 are present in the structure of every perception. However, a mind does not necessarily have perceptions of these components. Since perceiving a higher numbered component as a cause of lower numbered components requires being able to perceive the lower components, a mind cannot perceive the higher component until it has developed the ability to perceive the lower. Thus, a mind cannot "leap" to perceiving all components, but must develop through stages. A mind at a stage capable of a perception of a particular order (such as order 3) is able to perceive that component of its perceptions (such as component 3 "dependency") as well as lower components, but it is still not able to perceive the higher components. The development of higher orders of perception continues until order 4 (if it is ever achieved) where a mind can perceive the full structure, components 0 through 4, of its own perceptions.

At each stage, the mind stabilizes on an order of perception where the component perceived appears to be the most fundamental possible. A higher order perception (which shows that the component is not the most fundamental) is rejected as disturbance to the mind's sense of integrity. This rejection is accomplished by blocking tests which could allow the higher order perceptions to prevail. So the mind remains at a stage until higher order perceptions can be allowed without jeopardizing the sense of integrity.

Example: Seeing

As a way of exploring the different orders of perception at each stage, we consider sight as a form of perception and how a group of people see each other. In this example, there are people of different ages (and stages of development) and some toy balls.

We are going to examine sentences of the form "person 2 can perceive whether person 1 can perceive whether the object is there". When they are chained together like this, each "can perceive whether" is about perceiving another (lower numbered) component of perception, and so adds another order. This chaining can become quite verbose, but is necessary to rigorously show the structure of self-reference. Can this process of chaining one "can perceive whether" onto another go on forever? No, because there are a finite number of components of perception and so this "self-reference" of perception has a limit, at order 4.

Order 0: Baby

We start with a baby named, not surprisingly, "Baby." Baby can see the ball. (Baby can see the difference between when the ball is there or not, tracking the ball with its eyes.) In other words, Baby's mind has order 0 perceptions of its visual measurements while looking at the ball.

If a second baby, named "Baby 2," is looking at Baby, Baby might as well also be a ball because, while Baby 2 might see a round thing on Baby, Baby 2 doesn't see that it is an eye which points at the ball. Baby 2 can't see whether Baby is looking at the ball.

Order 1: Toddler

Toddler not only can see Baby, but Toddler can see whether Baby is looking at the ball. (Toddler can see the difference between when Baby is looking at the ball or not.) In other words, Toddler's mind has order 1 perceptions of the predictions connecting Baby and the ball.

Toddler can see whether Baby is looking at the ball. But Toddler cannot see whether Baby 2 can see whether Baby is looking at the ball. In fact, we know that Baby 2 can't do this, as explained above. But Toddler can't know this. To know this, Toddler would have to not only perceive predictions, but perceive that they come from theories (order 2 perception). Toddler could then look whether Baby 2 has a theory about whether Baby is looking at the ball and see that Baby 2 in fact doesn't have such a theory and is really just staring at Baby. So, when we say "Toddler can see whether Baby is looking at the ball," it merely means that Toddler can watch Baby's eyes track the ball, but not because Toddler has a theory like "Baby has the capacity of sight and is using it in regard to the ball."

Because Toddler doesn't have a theory of seeing, to Toddler it appears that everything can see in as much as Baby can see. To be precise, a mind at the stage of order 1 perception wrongly attributes what it perceives (ability to have order 0 perception) to everything, even objects such as a doll. In Toddler's experience, the world is a magical place where everything is alive. Entertainment for toddlers reflects this, where make believe chairs can talk and the trees have eyes.

Likewise, without a theory of the capacity for sight, Toddler doesn't have a self-perception as a perceiving subject which sees objects. Try to imagine how different a mind at this stage is. Without a theory of the mind as a perceiving subject, the mind does not connect its present state to previous states. Thus, it is easy to identify a mind a this stage. Toddler wanders disconnected from calendar time, in a perpetual present saturated with the gratifications of short-term cause/effect predictions of connections among sensory experiences. As an adult, the experience of fidgeting is a remnant of order 1 perceptions. E.g, the mind gratifies the prediction that running fingers along a piece of hair will come to the end of the hair, and repeats this over and over.

The predictions which constitute the world of experience with which Toddler identifies are produced by theories in Toddler's mind which Toddler doesn't have a means to perceive. But what happens if  Toddler briefly perceives one of these theories in operation? The short answer is that Toddler's mind would reject that perception as interference. Why? Because if the theory is seen (in an order 2 perception) as what produces the prediction with which Toddler is identified, then the theory has the possibility of supplanting the prediction, threatening the identification with the prediction and Toddler's sense of stability. Eventually, though, this is precisely what will happen in the transition to Kid's mind. Attachment to specific sensory predictions are given up because they can be obtained from the theories which were producing them all along, and stability of mind is now based on identification with the rich world of theories (beliefs, opinions, conceptual categories, etc.).

Order 2: Kid

Kid can see whether Toddler sees whether Baby sees the ball. (Kid can see whether Toddler is watching Baby look at the ball or if Toddler is just looking at Baby's head.) In other words, Kid's mind has order 2 perceptions of the theory which predicts that Baby sees the ball, and can attribute that theory to Toddler as "Toddler perceives that these conditions imply that Baby sees the ball".

Toddler thinks that the doll can see, but Kid knows it can't. So what? Isn't this just another fact about the world that Kid knows that Toddler doesn't? Why does Kid's ability deserve a different order of perception above Toddler's? Because higher orders of perception are about a mind perceiving more of its own ability to perceive, and Kid's mind can perceive that the doll doesn't see because Kid's mind can perceive that it itself sees. So Kid doesn't make Toddler's mistake of attributing order 0 perception to a doll. But does Kid make other mistakes?

Suppose another toddler, "Toddler 2," is looking at Toddler (who is looking at Baby). Kid can see whether Toddler sees whether Baby is looking at the ball, but Kid cannot see whether Toddler 2 can see whether Toddler sees whether Baby is looking at the ball. There are a lot of "sees whether" to keep track of in this sentence! We need some help, so we can simplify this to: "Kid cannot see whether Toddler 2 can see Toddler's theory about Baby". In fact, as we discussed above, we know that Toddler 2 cannot perceive theories. So as theorists, we can see whether Toddler 2 can see Toddler's theory, and the answer is no.  But Kid does not see this. Why? To see this, Kid's mind would not only perceive the world as theories, but perceive that it perceives the world as theories. Kid would then look whether Toddler perceives the world as theories and see that Toddler doesn't.

What are the implications of this for how Kid experiences the world? Because Kid's mind can't perceive the dependency of its theories, it appears that the theories should be self evident. Without an appreciation for the assumptions on which it is based, to Kid, whether or not someone understands a theory should be the same as whether or not they see that the ball is on the table. And while Kid experiences the flow of time and knows about events in world history, in Kid's mind the historical development of its concepts or even the fact that there was a time at which the concept didn't even exist is unclear. Kid is born into a world of theories, absorbing them intact, and it seems as if they could have been around forever.

Kid cannot take on hypotheticals, which helps to illustrate how a mind tends to "lock in" at a particular stage. The theories in Kid's mind depend on a set of contingent assumptions, and under other equally likely circumstances (such as being born in another culture) there would be different assumptions and different theories, beliefs, etc. But Kid cannot let go of the default set of theories to hypothetically take on others. It just seems evident to Kid that the default theories have to be the only true possibilities. (E.g. ask someone at this stage to try on the belief set of a different religion or nationality.)

The dependency of theory on assumption is a real fact about Kid's own mind, but what happens if Kid, even briefly, perceives this dependency (order 3 perception)? Kid's identity and sense of stability is based on the apparent solidity of the theories which make up the world view, so to perceive that a theory actually depends on a more fundamental assumption shows that the theory is not fixed, and Kid's mind rejects this perception as a disruption. A mind at this stage can typically resist the perception of dependency for an entire lifetime. The adult mind remains identified with order 2 perceptions, although these theories become limited down from those of a kid by the requirements of responsibility and survival. However, under various circumstances, a mind can shift to a stage based on order 3 perceptions.

Order 3: Theorist

From above, "Kid cannot see whether Toddler 2 can see Toddler's theory about Baby." But Theorist can see whether someone else can see a theory. I.e., Theorist can see whether Kid sees whether Toddler sees whether Baby sees the ball. (Theorist can see whether Kid is perceiving Toddler's theory about Baby or whether Kid is just watching Toddler look at Baby.) In other words, Theorist's mind has order 3 perceptions of the dependency of "seeing whether a person sees something" on attributing a theory of mind to that person, and Theorist can see whether another person is making this attribution.

Even though Kid "has" theories, Kid cannot perceive their dependencies and so cannot make and modify them. This is what Theorist does. Kid can be taught procedures for manipulating theories, but it is Theorist's understanding of the structure and dependencies of theories which allows Theorist to come up with the procedures in the first place which Kid follows. Whereas Kid resists perception of the dependency of cherished theories on fluid assumptions, Theorist embraces this perception, letting go of particular theories in favor of exploring the assumptions on which they depend, and on to deeper assumptions. To be at a stage based on order 3 perceptions, a mind doesn't just "come up with theories" but rather takes on different sets of assumptions experimentally, letting the theories which occupy the mind shift under the new dependencies. A Theorist may switch behavior patterns, not like the Kid would in following a theory ("if I follow these rules I can achieve this goal") but rather as a result of real-time perception of the ethical and strategic dependencies of actions as they are performed ("if I do this, it means I embrace these basic principles on which it depends, but do I really want to embrace them?").

A theorist's order 3 mind is the first order of perception which is able to reflect on its own processes in a useful way. So it is not surprising that the tectonic changes from earlier stages of mind are not remembered. The point of highlighting how the mind at earlier stages rejects higher order perceptions is to make it plausible to someone at the theorist stage that another stage is possible, and that the mind may not have perceived it so far because it is in the same process of rejecting perceptions which would show that the perceptions with which it identifies are not fundamental.

So what perception does Theorist reject? Theorist perceives that a mind develops theories over time. How does theorist perceive this mind on which perception depends? Theorist assumes the mind is a continuous entity (usually called "I") which persists through time and accumulates one perception after another. But even if perception did depend on such a persistent mind to accumulate them (an unwarranted assumption), this is experienced through a perception which activates in the moment. Can Theorist perceive this activation in the moment? No, Theorist rejects this perception. Why? Because this perception makes predictions which, if tested, would refute the experience of existing continuously over time. Without this experience, Theorist's mind assumes that it would stop perceiving (i.e., that Theorist would die). However, the ability to perceive does not depend on any perception that has to be fixated in the mind (as Theorist fixates on characteristics of a continual existence over time). This perception is maintained by blocking challenges to it, but a mind that permits those challenges can shift to a stage based on order 4 perceptions of the operations that maintain perceptions.

Order 4: Permitter

Theorist cannot see whether someone perceives a dependency on a persistent mind, since Theorist assumes that a persistent mind is necessary for any perception. But Permitter can see whether someone else can see that the mind doesn't depend on one of its own perceptions. Permitter's mind has order 4 perceptions of this process of enforcing the felt realness of a perception.

Even though Theorist can develop a theory (e.g., through academic research) that the realness someone feels of existing over time is a perception enforced by blocking challenges to it, Theorist cannot actually permit those challenges. This is what Permitter does.

Kid, at order 2, could be taught the concept of  "enforcing something seeming real by blocking challenges". So why can't Kid skip Theorist's order 3 stage and develop to order 4 to perceive "enforcement"? Because, without perceiving the order 3 generalized dependency of these perceptions, Kid would identify such a process with specific concepts from specific unquestioned traditions. This does happen often and results in dogmatic and inflexible thinking. Such a mind does not actually see "enforcement" in operation as Permitter does, but rather representations of this as allowed by a narrow conceptual framework.

Permitter has a different experience of time - not like Toddler who is isolated in the "now" through lack of the concept of other moments. Permitter is able to access memories and make strategic plans. But with order 4 perceptions, Permitter sees that such a timeline in which these events are located appears in a perception which momentarily activates. Thus, Permitter has the experience of "showing up" only during the time when the perception activates, not like Theorist of "showing up" over a stretch of time from the past into the present (on which a sense of continuous self identity depends). This allows Permitter to perceive how the experienced moment is "stitched together" (loosely speaking) from a mosaic of multiple constructed perceptions.

Finally, Permitter is at liberty to explore how mind perceives independently of needing any particular perception to be "in mind", including a perception like "I exist." By analogy, this is like exploring how photons function regardless of the pattern of light which they are forming, or studying the rules by which logic gates in a computer operate regardless of the program that the computer is running.


A perception uses components 0 through 4 as it tries and necessarily fails to characterize its environment. A mind develops through stages to be able to perceive each of these components in operation. At each stage, the perceived component appears to fully characterize the environment. In other words, the component is experienced as the "ground of reality". So to make this perception prevail, a mind blocks the tests which would allow perceptions of a higher order, since the higher order perceptions refute that the lower component is the "ground of reality". However, since perception of component 4 is perception of this blocking process and how it operates to enforce perception, a mind at this stage can "override" this blocking process "in principle."

Perception of component (as well as lower components)
Person at this stage
0: Measurement: Refutable assertions about what is present wrongly treats measurement observations as coming from nowhere
1: prediction: action causes unrefuted test (how measurement is made)
wrongly attributes its ability (component 1) to everything
2: theory: set of conditions implies certain predictions (how predictions are selected)
wrongly attributes component 2 (perceiving theories) to minds at order 1 stage (toddler) but not to minds at order 0
3: dependency: assumptions yield a certain theory (how theories are generated)
wrongly attributes component 3 (dependency) to holding a perception
4: Enforcement: Perceptions develop in the dynamic between blocking and permitting tests

back to home