The Material Process of Being in–and through–the Symbolic Matrix

This post is the next-to-last in my five-part series on Anthropology and Philosophy. Of course, the general theme of how these two disciplines have and may interconnect is rather open-ended. I’ve chosen to focus on particular issues that really strike me as highlighting:

1. parallel trends between disciplinary specializations and between the disciplines that could productively be drawn together–e.g., insights into the open-endedness and consequent, inevitable logical inconsistencies of both formal and everyday logical representation systems;
2. issues and insights from anthropological research that can benefit philosophical inquiry–e.g., Rappaport’s (1979) argument that in complex self-regulating systems, the most relevant Aristotelian level of causality is that of formal cause, underscoring the relationships among the system components; and
3. how the theme of biocultural connections can further contribute to a broader, interdisciplinary understanding of human existence and experience.

This post deals partly with the second theme. This is really a post about how the late British anthropologist Alfred Gell (1998) elegantly clarified and expanded the applicability of philosopher C.S. Peirce’s ideas to understanding human interaction with and agency in the world. Yet, this is more than just a summary of Gell’s own ideas, as they also inspire a more profound biocultural synthesis. Thus, this post also covers the more general–and admittedly ambitious–third theme, which is also the focus of a manuscript that I’m preparing for submission to a special issue of the open-access journal Frontiers in Psychology, dealing with “the future of embodied cognition.”

The main point of this post is that human linguistic communication is not only richer in embodied experience than we usually acknowledge, but it is also more thoroughly shaped and constrained by symbols and their indexical interconnections than scientific accounts of human social interaction usually assume.

Symbols in Control

As discussed in Anthropology and Philosophy III–and as will be very familiar to some readers–C.S. Peirce developed a particularly influential, general analysis of logical representation systems. In this analysis of “logic as semiotic,” Peirce argued that we can–and do–communicate with different kinds of signs, which have different formal properties and patterns of association with the objects they refer to.

Arguably the most important way that logical signs vary is according to Peirce’s trichotomy of indices, icons, and symbols.

An index signifies a particular relationship between phenomena constituting some state of affairs, so that the index phenomenally implies or is connected to its object. More simply, an index conceptually points to its referent. When we cognitively experience or produce an index, its object is invariably connected to it in a logically consistent way. To use a particularly well-worn example, we may perceive or call attention to smoke as an indexical signifier of fire.

An icon has an arbitrarily simplified resemblance of its object, and as such, its own form signifies or evokes the object, even if we don’t experience that object directly.

A symbol is largely or entirely arbitrary with respect to what it represents. Whereas the icon is more conceptually complex than the index in its capacity to evoke something absent or lost, the symbol adds an additional, extraordinary level of conceptual complexity. The symbol is nested hierarchically alongside what it represents, so that–in turn–the whole symbol-referent concept can focus our attention on an object–on a state of affairs in the world.

Indices and icons–once learned–cognitively simplify and organize our perceptions and expectations.

The symbol–once learned–can cognitively facilitate understanding, navigation, and prediction about the environment.

In fact, in communication symbols are particularly efficient social tools for pointing toward perceived states of affairs in the environment. The primatologist Michael Tomasello (2008) has persuasively argued that much of human interpersonal linguistic communication is about managing and influencing joint attention in dyads or groups.

Yet, as practical tools for taking in information and acting in the world, symbols are impossible to fully control. In fact, sometimes they control us.

I have heard the following story from more than one social scientist, recounting efforts to learn the local language from people native to very remote, non-literate communities. The social scientist would point toward something and use his/her limited local language skills to ask, “What do you call that?” The social scientist would begin to hear the same word, regardless of what it was he/she pointed toward. Slowly, it dawned on the researcher that he/she was hearing the local word for finger. Pointing one’s index finger is not a “Peircean” index signifying what one is pointing toward. It is an arbitrary symbol.

This suggests that symbols have a dual, contradictory nature.

And which side we experience depends on our embodied and social mastery of shared symbol-referent concepts in a community.

On the one hand, symbols are embodied social tools that give us control over our environment. When combined with the direction of the sender’s sustained, steady visual gaze–which is an index of the sender’s joint perception and communicative intent–symbols for noun or pronoun referents give the audience very specific information about what to expect if they train their gaze in the same direction. But this is only the case if the arbitrary symbol-referent relationship is something that the sender and audience have both learned.

On the other hand, the arbitrariness of the symbol’s form can allow the symbol to control us.

Consider any instance in which you have said or written something. You consider it to be an intentional act. We usually don’t consider such intentional behaviors to make something–in this case, an auditory or visually transmitted message–that can actually, directly come back to influence us.

Let’s be clear. I’m not talking about when you say something that ends up having consequences for you because someone else reacts to it.

I’m talking about the symbols themselves coming back to influence you … sometimes immediately.

The figure above illustrates why it is useful to consider symbols as controlling us in some ways, although we control symbols–and what they refer to–in others. We will come back to this point below. For now, it is worth noting that once we start to think in terms of control or constraint, we begin to see just how much agency that symbols have as constituents of our environment. This is a basic insight of Alfred Gell’s (1998) work on Art and Agency.

Recursively Constructing Symbols from Symbols is Like Riding a Bike

The fact that symbols can control us is especially important because we perceive and seek to influence our environment through particularly complex, symbolically constituted and recursively constructed symbolic representations. In other words, our representations are symbols that consist of a bunch of other symbols and their mutual arrangements. Symbols are so ubiquitous in our environment that it would seem to be a good idea to try to understand them–and rather thoroughly, at that.

The fundamental concept of the symbol is constituted by the arbitrary association of the symbol to its referent. And it is this holistic symbol-referent (as Peirce technically called it, “representamen-object via interpretant,” or as the French structural linguist Fernand de Saussure called it, “signifier-signified”) relationship that directs, diverts, or focuses our attention to some object–that is, some state of affairs in the world–as we use symbols to construct thoughts and utterances.

So, when someone utters the sentence, “There’s a beautiful tree over there,” it is the mental association of the complex phonological symbol “tree” with our memories, experiences, and other concepts of trees or “treeness” that allows us to understand the speaker’s intention. We are to visually scan for a tree that the speaker implicitly asserts as remarkably beautiful … and within or near our joint field of sight.

And there’s nothing necessarily intrinsic to the auditory, visual, or tactile symbols we use that indexically connects them to their referents … except for the actual, socially (or occasionally, privately) repeated coupling of the symbol, its referent, and the symbol-referent’s experienced-as-relevant objects in the world.

Thus, as in the figure above right, what does the visual or auditory form of the English phrase “the larch” have to do with larches, trees, forests, or lumberjacks? Polish subtitle and dubbing translators would know that the Monty Python scene about the larch is about the tree known as “modrzew” (genus Larix), species of which can be found in the mid-upper latitudes across most of the Northern Hemisphere (see botanical illustration below left).

English speakers cognitively process the phrase “the larch” in association with some learned, remembered concept(s) relevant to larches. Consequently, when larches or similar trees stand out in the setting of a particular experience that we participate in, observe, have related to us linguistically or artistically, or imagine–well, it is the linguistic symbol “larch” that is evoked.

Peirce’s argument about symbols (along with the similar, nearly contemporaneous work by Fernand de Saussure, one of the founders of modern linguistics) is important for all of us to consider. Usually, once we thoroughly learn a symbol-referent concept, we experience the symbol as naturally, directly connected to objective states of affairs that it evokes or focuses our attention to. What Peirce realized is that symbols are formal, logically structured things that we can make, perceive, or imagine … and they logically require an evocative concept–or network of concepts–that mediates and filters our comprehension of when that symbol can help us to make sense of, communicate about, and take action to change states of affairs in the world.

The most important constraint on the symbol’s actual form is that it has to be sufficiently distinct–in its auditory or visual or tactile form–from other symbols that are arranged to constitute a particular message. Of course, a given word can have multiple meanings–that is, multiple unrelated symbol-referent relationships. The English word “mean” can refer to concepts of arithmetic average (noun), denote or connote (verb), malevolent (adjective), or common, vulgar (adjective). But in practice, the word’s formal contrasts with adjacent, grammatically-different-but-logically-related symbols can provide us with sufficient information about which meaning of “mean” is relevant.

“He is mean” can thus signify a clear message, with a largely mutually shared conceptual web of referents and an unambigous real-world object–for speaker and audience alike. On a linguistic level, this is because “mean” is hierarchically nested within a larger, syntactically structured symbolic utterance. In this example, “mean” is:

1. part of a longer string of phonemic symbols, which follow an order constituting a nearly continuous sequence of sound variations, and
2. part of  a sequence of morphemic symbols (that is, words, which are made up of subsequences of phonemes), themselves constitute an ordered pattern of syntactical relationships.

One of the remarkable lessons to draw from Peirce’s (and, in this case, also Saussure’s) insights is this:

When we practically, fluently use language in everyday interaction and thought, we unconsciously construct and decode linguistic messages by extraordinarily rapidly moving back and forth between non-nested hierarchical levels of symbolic organization.

• Before formulating the utterance “he is mean,” we have to make a determination (usually a snap, unconscious one, at that) that concepts of malevolence, of self and the group, of past and future, of webs of social connections, and relevant practices of social judgment are really pertinent to the social situation or set of private concerns, interests, and reflections at hand.
• We then put together information about gender, along with the immediate social and spatial factors that suggest a pronoun (“he”) works better than a proper noun (“Bob”).
• We further consider thoroughly learned, embodied knowledge of hierarchically organized symbols, from the phonemic (constituent sound symbols), to the morphemic (basically, word symbols), and on to the syntactic levels. Syntactic patterns in the language at hand will have a particularly strong non-nested hierarchical filtering influence on how we cognitively work through the relevant symbol-referent concepts in formulating or decoding a linguistic message. Understanding how the whole symbolic utterance is syntactically structured into logically interconnected parts helps us put together the sounds in an order that will signify the (presumably socially relevant) concept “he is mean.”

To put all of this in Peirce’s own terms, symbols are part of our unconscious habits, but they can also be pragmatic tools. In all instances, symbols aren’t just about thinking. They are embodied instruments for action. As such, learning symbol-referent concepts and recursive, hierarchical symbol-symbol concepts is formally similar to learning how to ride a bike.

We experience embodied processes of learning to combine simpler, already-learned skills into more complex motor or cognitive behaviors in many contexts. Not just when we produce or interpret a linguistic utterance. Such complex learned behavioral sequences include, of course, how to ride a bike. Or how to make a jump shot in basketball; how to solve integrals in calculus; how to knap a flint handaxe; or even how–for the infant–to begin to sit up, speak or walk.

Overall linguistic competence is just one system of learned, embodied mastery–mastery of rapidly making sense of a complex goal and putting together relevant parts into a seamless sequence to achieve that goal.

Much cognitive research over the past several decades has been devoted to exploring embodied perception and action–by humans, non-human animals, and even robots–in the environment (Chemero 2009; Clark 1998; Gibson 1979; Pezzulo et al. 2008; Rowlands 2008). Complex behavioral sequences may require a steep learning curve, because the component parts must be mastered along with the sequence in which those parts are carried out. But it is also clear that, once we have learned the parts and their sequences, effective sequence-dependent relationships between parts actually leverage information or energy from earlier steps, allowing us to focus on the end goal or the overall form of the whole.

In fact, learning the parts and their sequences is not a strictly nested hierarchical process. Imagine a complicated task that may be reasonably divided into many parts. If we alternately work on learning and practicing the first two parts, A and B, with the symbolic knowledge that they will ultimately be put together in the sequence A→B–and that A and B alone don’t do us much good toward reaching any relevant goal we have–then A becomes an index of B. Cognitively, A can evoke B … as long as we have the cognitive processing power to practically, simultaneously carry out A as we begin to imagine B and the transition from A to B. Simply experiencing the practical transition from A to B during learning can help us focus and motivate: (1) further learning toward the overall goal and sequence; (2) more effective learning of A and B and the transition between them; and (3) embodied understanding of how it’s relevant for B to follow A, and not vice versa.

In other words, symbolic representation of the whole sequence and its relevance to a goal actually makes it more efficient to learn the individual parts, subsequences of the whole sequence, and why the whole is relevant to the goal. This efficiency is gained because the symbolic representation itself serves as a simple, indexical way-marker, cognitively reminding us that our learning is keeping us on the right track. When we’re unsure or confused or frustrated, we don’t have to stop and devote all of our energy to reflectively, consciously thinking through the entire sequence in detail.

As learning progresses from from bumbling frustration to unreflecting, masterful habit, we progressively construct an indexically ordered embodied representation of the whole activity and its parts. This representation is often unconscious, at least in part. More interestingly, its evocative indexical structure gives the representation a quasi-holographic property. Thinking of pedalling a bike evokes other parts of the activity, along with the whole.
Whether we are talking about bike-riding or we are talking about talking, indexically structured and highly evocative embodied representations shape our cognition and behavior throughout our lives.

Symbols as Indices: Symbols Influence the Focus of Our Attention

This perspective on symbolic representation as a catalyst for complex learning helps us to understand that the human linguistic capacity is not just–or even mainly–about consciousness … about the capacity for reflection and self-awareness. Language indeed allows us to think intentionally about long-term goals, counter-factual scenarios, wished-for possibilities. But as Gell (1998) implied in Art and Agency, linguistic representations–along with icons and intentional, ambiguously iconic works of art whose symbolic referents are also unclear–are not just symbols. They are also indices, potentially with multiple objects, pointing our attention to part-part and part-whole relationships in real-time, complex learning and action sequences. Language thus unconsciously pervades and aids–but also channels and constrains–our learning and imitation capacities.

And this brings me to the other remarkable lesson that may be drawn from Peirce’s Index-Icon-Symbol trichotomy, one :

Symbols are not just symbols. They also dynamically function as indices, guiding us cognitively to related symbols or referential concepts.

Whereas chimpanzees and bonobos have learned symbol-reference concepts in captivity, only humans are naturally, ecologically dependent on social production of symbol-reference concepts. We continuously structure and restructure systems of symbol-reference concepts and their practical interrelationships into grammatical languages. As Deacon (1998) insightfully points out, the fact that at least some members of those ape species most closely related to us can learn symbol-reference relationships in captive laboratory settings–without using them naturally in the wild–suggests that human symbolic thought and communication is really a (not-completely-nested) hierarchically organized system of indexically interlinked–and often, iconically formed–sign concepts.

Indeed, there’s a surprising amount of conditioned, associative learning involved in developing iconic and indexical recognition or production abilities (Deacon 1998: Chapter 3). Another way of making this point is that Peirce developed the icon-index-symbol analytical trichotomy to understand signs, which humans can both think and use to communicate (send, receive, and interpret) messages–messages that signify something about a state of affairs in the world. But Deacon emphasizes that the icon and index concepts are helpful for taking an evolutionary view of the origins of cognitive abstraction, memory construction, and learning. Because if arbitrary symbols can indexically signify other symbols, learned iconic representations–which are abstractions of highly complicated (possibly multi-modal) sensory perceptions–can indexically signify other icons. As suggested by the two successive figures below-right, Peirce’s basic Index-Icon-Symbol trichotomy simplifies our own representation of a constructionist model of cognition and learning.

Asymmetric Indices and the Unfamiliar

I have discussed in earlier posts how symbols and their relationships influence our awareness of the world–in the form of marked-unmarked structures. Markedness is something that can particularly strongly shape larger patterns of political relationships. It does so by differentially evoking otherwise-ideologically-masked unequal relationships between individuals and groups; the differential evocation depends on whether we focus on the party with greater or lesser institutionalized power. To put this in Peirce’s terms, some environments support the construction of cultural, embodied representations that are strongly asymmetrically evocative. A marked-unmarked relationship between symbols (or icons, for that matter) involves A being an index of B, but not vice versa.

Now, it’s important to note that I am following Gell (1998) and Deacon (1998) in their independently derived, interestingly similar interpretation of Peirce’s index concept in the context of embodied cognition. They view indices as structuring of conceptual relationships among symbols or their parts, icons, and objects.

The indexical configuration of relationships among signs actually entails that signs–be they icons or symbols–have a dual identity.

On the one hand, the icon or symbol will refer to an object (by formal similarity or convention, respectively).

On the other hand, these signs become:

• indices of their objects;
• indices of other signs they evoke or stand in close relation to (in hierarchically constructed, more complex representations); or
• objects of other signs’ indexical identities

Linguists who have sought to synthesize Roman Jacobson’s ideas about markedness with Peirce’s icon-index-symbol trichotomy describe the marked-unmarked relationship between two symbols as being a special category of sign: that of the cognitively mediating sign, or interpretant. Peirce suggested that the interpretant is the conceptual sign that links symbols to their objects. In this case, the marked-unmarked interpretant-concept signifies an asymmetrically evocative relationship between two symbols (Andrews 1990). However, when Peirce posited the existence of an interpretant concept, arguing that it must mediate between a symbol and its referent object, what he was describing most often takes the form of the symbol doubling as an index that signifies something else. This is particularly the case in processes of evocation and recursive representation construction.

Consider a hypothetical symbol, S1. It is logically linked to object O1 by social convention. Symbol S1 is logically distinguished from S2 and S3, etc., as much by the formal differences among them, as by syntactic or other referential differences among their objects, O1-O3. And if O1 and O2 represent asymmetrically interrelated phenomena (for instance, racial or ethnic political domination of one identity group over the other), then the corresponding symbols S1 and S2 may stand in a marked-unmarked, asymmetrically indexical relationship. Logically, if O1 is an index of O2, but O2 is not an index of O1, then an instance of O2 will evoke S2. And a representation of S2 will index and evoke O2. But an instance of O1 or representation of S1 will evoke O2 and S2.

Although I have emphasized the political dimensions of marked-unmarked symbolic relationships–say, between social identities–there is a much much more fundamental and pervasive kind of marked-unmarked phenomenon. Something we experience as unfamiliar is marked in relationship to the familiar. The familiar will already involve iconic–and possibly, symbolic–representations in indexical relationships. However, the unfamiliar will involve the formation of new iconic representations in search of indexical connections to other iconic or symbolic representations.The unfamiliar throws us out of symbolic homeostasis.

I emphasize that iconic representations of sufficiently unfamiliar experiences are thoroughly embodied (Clark 2008; Rowlands 2006). They are shaped by a complex interplay among direct external stimulus reception, emotional response circuits, more general feelings or awareness of embodied disequilibrium (ranging from physiological sensations–such as heightened alertness or relaxation, internal pain or discomfort–to feelings that indexically point to general emotional states, such as happiness, contentment, fear, dread, desire, jealousy, etc), and evoked iconic memories or rapidly constructed symbolic representations.

Bodily disequilibrium is basicially a state that uses already-available, limited metabolic resources inefficiently or causes maladaptive behavioral responses in the environment. It can arise from three possible factors, each of which can dominate or work together. One is a highly rapid or large-scale change in the extrasomatic environment, overwhelming normal homeostatic capacities. Examples here range from injury and illness that reduce alertness and homeostatic bodily interaction with the environment to stimulating situations–such as predator threats, prey-capture opportunities, or agonistic social interactions–that raise alertness or focus, involving sustained or intense interaction. Although the ability for an animal to shift temporarily into such a disequilibrium state is adaptive for surviving over a longer life-course in a complex environment, the inability to return to a homeostatic range with less energy and nutrient turnover is certainly maladaptive. I would argue that the human embodied cognitive process of dealing with the indexically marking unfamiliar evokes the unmarked familiar, in order to return to an overall physiological, emotional, and cognitive homeostatic range.

It is in this sense that symbols most often control us.

UPDATE NOTE (19 January 2014): I have corrected the citation to Gell’s (1998) Art and Agency, adding it to the reference list below. Considering how important Gell’s book is for inspiring this post, I regret the mistake and can only suggest that initial errors (inaccuracy and omission) reflect my enthusiasm and anxiety to get the ideas out there. I have also modified the second-to-last paragraph above (now in italics), concerning disequilibrium and homeostasis.

References

Andrews, E. (1990). Markedness Theory. Duke University Press.

Chemero, A. (2009). Radical embodied cognitive science. Cambridge, Mass.: MIT Press.

Clark, A. (2008). Supersizing the Mind : Embodiment, Action, and Cognitive Extension: Embodiment, Action, and Cognitive Extension. Oxford University Press.

Deacon, T. W. (1998). The Symbolic Species: The Co-evolution of Language and the Brain. W. W. Norton & Company.

Gell, A. (1998). Art and Agency: An Anthropological Theory. Clarendon Press.

Gibson, J. J. (2013 [1979]). The Ecological Approach To Visual Perception. Psychology Press.

Pezzulo, G. (2008). Coordinating with the Future: The Anticipatory Nature of Representation. Minds and Machines, 18(2), 179–225. doi:10.1007/s11023-008-9095-5

Rappaport, R. A. (1979). Ecology, Meaning, and Religion. North Atlantic Books.

Rowlands. (2006). Body Language. MIT Press.

Tomasello, M. (2008). Origins of Human Communication. Cambridge, MA, USA: MIT Press.