Those theories opposed to cognitivism do not seem to have gotten rid of any of the dichotomies and, instead, put them closer together. Consider Figure 5 below. Imagine the figure lined up with Figure 1 so that the draftsman lines up with “representation” and the nude lines up with “reality.” The “representation” column stands for a way of considering the draftsman. The “reality” column stands for a way of considering the nude. An arrow is equivalent to a process. A circle is equivalent to a result. Each row then becomes an
“interpretation” of the relationship between draftsman and nude. Row A illustrates the traditional dichotomy between representation and represented. The process (arrow) terminates in a result (left circle). Together, the two correspond (the arc) to what goes on in reality. This is an indirect perception or representational view of phenomena. It is the result of the process (left circle) that is correlated with the circle in reality (right circle). Below, in row B, the process (arrow) is directly connected with the result (right circle). This illustrates direct perception, as advocated by Gibson (1986). We connect with things directly because the process is not “in here” but “out there.”
155
Figure 11 (5 in article). Direct and indirect perception; direct and indirect instruction
As stressed in the example of the draftsman, what is made to “stand in” for reality (a body, a brain, an image in that brain, firing neurons, a map, or a text) is strictly speaking also simply an object in the world. In order for a dichotomy to be in place, a process has to be implied by the object. This was accounted for in the process/pattern dichotomy. What was not examined is that – unlike the very visible arrows in Figure 5 – the process is not present at any time in the case of the draftsman/brain state/text. It is continually implied by the given object. A proper visualization is more closely related to recipes and IKEA instructions that signify a process by step-by-step depictions of a state (see Appendix I). In order to clarify this, we have to differentiate between two forms of instructions – “direct” and “indirect.”
Carstensen and Nielsen (2001) provide a helpful example of the difference between the two:
At the end of a berthing maneuver the master may announce over the walkie-talkie that ‘we are in position’ or he may say ‘tie her up.’ For all practical purposes both utterances will lead his men to perform the same set of tasks, namely finalizing the mooring procedure. The first statement signifies a state of affairs within the field of work, while the second statement is a direct specification of the task to be performed.
(p. 88)
The direct instructions specify tasks, procedures, and methods by depicting each of the states. In the case of the karate kata (a series of movements to be executed by a karate practitioner) (Appendix I), each state depicts either a beginning, end, or resting point of a movement. Movements done in short succession are grouped together. Movements that may be ambiguous from one angle are shown from a different position. The progression through the movements are indicated by a straight fully drawn line while dotted arrows, minor arrows, and fat blocks are used to indicate direction of movement, specific limbs moving, and duration of stances. The entire process is numbered, and a nomenclature is provided.
Various other signs signify additional information such as when (and what) to exclaim. The focus of attention is of course the many drawings of a body in different positions. Although the entire diagram does its utmost to make the process intelligible, what it cannot provide is actual movement. As Ingold (2000) has convincingly shown, converting (verbal or graphic) instructions into actual bodily movements is immensely difficult. Ingold uses the example of knot making to argue that “practices… cannot be understood as the output of any kind of programme” (p. 358). In Figure 5, the direct instructions are depicted in row C. The horizontal line signifies a process. The vertical lines at each end signify start and end states.
156 If we take the figure to represent a brain, the line may stand for synapses firing and cognitive processes unfolding, but it has to make do with an object (a static image) (Roepstorff, 2009). The image of a brain can signify a starting or end state, but it can only depict a state, not a process. In representational accounts, the end state or the entire process is interchangeable with the representational result (a perception, a mapping).
The examples of indirect instructions may seem easier (Appendix I). They depict how to put together a Lego set and to tie a tie, respectively. The indirect instructions “signifies a state of affairs within the field of work” – that is, instead of depictions of each state relative to the body, our attention is limited to what is in front of us. As such, the indirect instruction corresponds to a first person view (or what you see when you enter – process – a third person perspective as I Figure 2) while the direct instruction is equivalent to a third person perspective. Part of the ease of assembling the Lego set may be ascribed to a certain structuring (or built-in constraints) of how the pieces are able to fit together. However, as in the case of tying knots, the sense of what is going on in the instructions is only obvious once the movements have already been accomplished. These instructions are only clear and concise to the person who has already put together Lego sets or tied a tie. Everyone else will face a task of deciphering the instructions and turning them into actual movement. Examples of elements that are obvious only because we are accustomed to them are the sequential ordering of steps and the assumption that it is the same object at different moments. In Figure 5, the enactive view is signified by row D where the process connects directly with the object in question. Indirect instruction relies on this connection. Seeing observable phenomena as cognitive processes requires that a pattern (like the knitting pattern) is recognized as a transform of an algorithm (the process). Gibson’s (1986) affordances (direct perception) can be seen as indirect cognitive instructions (e.g., that something is jump-over-able can be seen directly because what is seen (object) is “seen” by a sensorimotor process that does not see the object as much as it sees a transform of a movement).99 Crucially, however, we can note that indirect instructions also rely on objects (static images) to suggest a process.
To be clear, when both direct and indirect instructions are offered, we do not come to the conclusion that we must try to do two things. Similarly, when we consider both the process and the result, we ought not to think that we are comparing two phenomena. We are, in fact, considering the same phenomenon in two different ways. These two ways are commutable or are transforms of each other.
For anti-cognitivist theories, the departure from traditional dichotomies (row A) lies in recognizing that perception, for example, is direct, not indirect (row B), and thereby transgresses the divide between a subjective and an objective domain. In the above analysis, this was shown to rely on redundancy between process and pattern, that is, that the two are interchangeable or commutable. What Varela (1992) seems to be getting at when he states
99 The idea is similar to that of action-oriented representations that simultaneously describe aspects of the world and prescribe possible actions (Clark, 1998, p. 49), only that instead of a map acting as a controller, the world itself acts as a controller, i.e., the world is its own best representation (Brooks, 1990).
157 that subject and object arise together is the fact that we can identify two different phenomena. What is unclear is why we should do so. Interpreting process as instruction shows that the commutability of process and pattern lies in such a way that instructions can be derived from both proxies (the subject and the object). Row C thus provides direct instructions by taking a third person perspective on the process via an object, e.g., a brain, while row D provides indirect instruction by taking a first person perspective on the process via an object, e.g., a knitting pattern. If these instructions happen to be the same, then we can reverse the logical order for a more economical account. The same instructions/process can be captured in a phenomenon, that is, an object or static image in two ways. The way in which we capture it depends on whether we have the process or the pattern in mind (e.g., subject and object). This aligns with the use of the term “direct” in Figure 5. In perception, for example, it is clearly “what we perceive” that is of interest. Perception is thus seen as direct when the process terminates in and with the object (pattern) while it is seen as indirect when it terminates in a representation (process). This is directly mirrored when we are methodologically interested in the process. Instructions are direct when signified by the body or the subject (process) while indirect when signified by the pattern (object). If this reversal stands, then processes are not derived from phenomena; rather, phenomena express processes.
The vis-à-vis of subject and object is therefore valid when we compare and side-order two commutable versions of instructions (process) but invalid in terms of examining separate and distinct phenomena (patterns) that have nothing intrinsically to do with each other.100 Examining the phenomenon of a brain will tell us nothing about knitting patterns unless we inspect brain images with respect to such patterns. The “error” committed in the correspondence problem and similar dichotomous problems is that we oppose interchangeable phenomena that stand proxy for the same process with respect to each other, but what we subsequently compare are their indistinguishable cousins, that is, the patterns we find when we examine each separately.
Put in more colloquial terms: What you see (the pattern) is a transform of how you see (the process). It is possible to speak of an ongoing, situated, provisional accomplishment (Barley, 1986) – but one has to keep in mind that there is not something (thing or object) being upheld by a process. Rather, what is seen is a pattern that takes on its peculiar – varying and differentiating – aspects. A thing is not a fait accompli but shifts and changes.101 There is not a thing first that is then subject to a process of perception. A thing is
100 A great part of the error lies in the idea that things and processes have to “take place,” that is, we have to situate a phenomenon somewhere. Once a phenomenon is in its place, we proceed to relate it to phenomena that are located in other places. Insofar as there is a process going on (e.g., imaginary, perceptual), it has to take place somewhere – and that somewhere then becomes the head. The result of the process (e.g., the phenomenon perceived) is located outside the head.
101 This is not a claim that there exists nothing stable. The point is intended to counter the idea that a process should be about a thing or that a thing is a pattern (out there). Rather, the things we see are what we see while in the process of patterning – they are patternings.
158 patterning (a process) as opposed to seeing a pattern.102 We may attempt to capture the process in patterning and find an algorithm (process) or establish an object which stands proxy for it, but this process – in the same line of thought – is not a process in a medium that carries content or information about an object; it is (for instance) the very sensorimotor process that creates the pattern. Changes and shifts in processing are equally responsible for shifts in the pattern.
The fundamental error arises as a result of the particular task of capturing a phenomenon (instructions) as an object or static image (a thing). Depending on whether we want to capture the patterning as pattern or process, we lean towards one of the two sides of a dichotomy (two different ways of instructing). We are so accustomed to converting pattern into process and vice versa that we can easily see the body of a man and from that alone assume his perspective and “see what he sees” (Figure 2). The connection between the body or brain of the perceiver with the perceived is therefore obvious and taken together, but it is utterly confusing once we begin to compare them as patterns. Comparing patterns assumes comparability in the very act of juxtapositioning. That is why we struggle when presented with Borges’ (1999) Chinese taxonomy. From entry to entry, we are at a loss to find commonalities, and even if we find one between two entries, it is immediately undermined by the next entry. The error can be likened to a faulty recursive operation. A recursion is a mathematical function that can solve complex problems if used properly. It does so by breaking up the problem into smaller sub-problems of the same type as the original, which are then solved and combined for a final result. The tricky thing is that every sub-problem is examined for part of an answer to the problem. If it finds an answer, it stores this part. If it does not find an answer, it repeats the process of breaking up the sub-problem into smaller sub- sub-problems – again, of the same type as the original – and so on and so forth. Finding out whether a problem is answered or should be further broken down depends on an exit condition (or base case). If the answer fits the exit condition, the (sub-)problem is solved.
Once all problems are solved, all the answers can be compiled into a complete answer. This type of problem-solving is an everyday occurrence.103
The error is generated by using a particular format (thing/object/static image) to capture a dynamic phenomenon (process) – that is, we try to use words or images to capture a particular way of seeing (instructions). In the case of the correspondence problem, we are able to return an answer precisely two times, that is, as an object (direct perception/indirect
102 This is not identical to the ontological claim that all things are processes (see Seibt, 2013) although it has many overlapping implications. Stabilizing patterning and calling it “a process” does not make it less of an entity. Process philosophy thus remains tied to a particular “thing” format, e.g., dismissing “reality as an assembly of static individuals” but embracing the proposition that “the world is an assembly of physical, organic, social, and cognitive processes” (Seibt, 2013) or distinguishing between “‘activity-like,’ non-developmental processes or ‘goings-on’” and “‘accomplishment-like,’ non-developmental processes or ‘comings-about’” (Seibt, 2013); each of these can be likened to what has been termed process and pattern in this article.
103 Herbert Simon has, for instance, described the design process as a decomposition of an overall problem:
“The whole design, then, begins to acquire structure by being decomposed into various problems of component design, and by evoking, as the design progresses, all kinds of requirements to be applied in testing the design of its components” (Simon, 1973, p. 190).
159 instructions) and as a subject (indirect perception/direct instructions).104 We then proceed to compile them for a complete answer. Both answers are correctly formatted (objects) and are computable (instructions can be carried out), but we will never be able to compile them properly because they are twice-captured commutable instructions. It is this compiling that we attempt when we speak of internal processes in the brain and then ask how they correspond with external objects or vice versa. The earlier-mentioned “redundancy between different modes of operation” is the redundancy between direct and indirect instructions.
Compiling two sets of instructions is not quite like asking what the relationship is between Sagarmatha and Chomolungman (both names for Mount Everest). Rather, it is more like securing your seatbelt and then asking where the buckle goes.
This explanation brings us back to the modalities with which I opened the article. It was stated that every mode employs a particular format of what is accepted or interpreted as an event. The examples of a speech act, a move, an utterance, a deal, or a sketch were given.
The particular modality under examination here is the linguistic mode (writing). The format employed in the linguistic mode is the particular way a phenomenon is captured, that is, the linguistic mode captures phenomena by seeing and naming things. We have seen how a thing can either signify a process or a pattern, but it is closer to the mark to say that we use the format of a thing to capture a phenomenon as either. The discussion has been focused on discussing how rules (processes) generate structure (pattern) and, conversely, how rules (processes) can be construed from structure (pattern) to sort out how dichotomies arise. The format of a “thing” itself has mostly been referred to in a self-evident or commonsensical manner. However, the format itself carries the key to the question of why dichotomies are so resilient and pervasive:
If what you see is a transform of how you see, then a thing is not something we perceive. A thing is the answer to how we break up the world into “see-able” bits, or to use a computer term, it is how we render the world. Things are how we see the world (not what we see in the world). “The thing” is therefore a format that simultaneously answers the questions
“what is?” and “how is it?” The format acts by the law of the instrument, meaning that, because it is the accepted format of when something takes place, then “everything that takes place” takes place in that format – expressed in colloquial terms, “to a hammer, everything looks like a nail” – but taken in reverse: everything that a hammer nails assumes the format of a nail. What has been unveiled is that when a hammer turns upon itself, it will not see a hammer, but a nail (e.g., when we turn to see ourselves, we are satisfied with the answer that we are “subjects,” that is, a type of thing). It has also been shown that from things (patterns), it is always possible to infer the processes that gave rise to them (we can reverse engineer the processes from the patterns, so to speak). A dichotomy is therefore a setting of the subject (a thing signifying a process) in opposition with an object (a thing signifying the resulting pattern). Both of these can be turned into the other. The process can be computed
104 The astute reader will note that these are in fact identical; only the “topic” or phenomenon in question differs. This means that one can easily misconstrue direct instructions as working because they directly address
“what to do.” However, the process at work in direct instructions is not the one depicted in the images or described in the text; it is the process that enables one to see or understand the pattern at which the process is directed. The fact that this understanding can go through a different pattern (indirect) is a matter of translation.
160 to yield the resulting pattern, and from the pattern, the process can be reconstructed.
Insisting that there are two different phenomena in a dichotomous relationship is like insisting that the phenomenon of baking a cake holds a fundamental opposition between the baking and the cake.105 The opposition thus arises from the attempt to capture the same phenomenon as both process and pattern, which in the format of things requires two things.
The format is extremely powerful. We expect the world to reveal itself in the form of things.
Even when nothing has been perceived and no event has taken place, the format is still in place as an implicit assumption about the (past, present, or future) form that any phenomenon can take. Any event or thing yet to materialize is already pre-rendered because we know it has to appear in this format. A format thus installs itself, like a Trojan horse,
Even when nothing has been perceived and no event has taken place, the format is still in place as an implicit assumption about the (past, present, or future) form that any phenomenon can take. Any event or thing yet to materialize is already pre-rendered because we know it has to appear in this format. A format thus installs itself, like a Trojan horse,