CHAPTER ELEVEN

CULTURAL SYSTEMS

by Hugh M. Lewis

Cultural systems consist of partially closed, semi-autonomous systems of information that are primarily symbolic in its form of integration. These are rooted in the human capacity for symbolization that is hard wired to the brain, to which language is integral as a form of mediation, and that involves and integrates facets of behavior, motivation, perception, imagery, concepts, rationalizations, social relations in a holistic and coherent manner.

Cultural systems thus are implicitly rule-determined systems, and this implicit normative structure of cultural orientations entails the possibility that its order can be propositionally described and represented, especially through computer simulations. The cultural representation of symbolic knowledge in terms of a cultural system provides a viable alternative to artificial intelligence research and cognitive science, one that has not yet been fully realized, and is aided by the fact that cultural knowledge appears to be domain specific, and available to systematic forms of response elicitation and rule formulation.

Computer simulation of such knowledge entails the use of entropic information functions in the formulation of decision rules which allow the successful navigation of certain discrimination networks. Networks that are constructed from the elicitations of representative cultural samples. The culturally defined pathways through these networks are predictable in terms of the decision rules. There is a certain amount of parametric learning which can be included in a feedback loop between the searching of alternative pathways through such networks, the construction of decision trees representing these pathways and the formulation of rules which lead to the successful navigation through the networks.

A wide variety of different kinds of cultural knowledge, at different levels of analysis, are thus able to be represented and simulated within the shell of a single such computer based cultural system. What is not understood is the binding problem of the interrelationship of the different domains of knowledge and symbolization to one another--and it is at this point that alternative theoretical interpretation enters into play.

It appears, based upon previous research, that an quaternary structure of analogical association underlies the organization of many different forms of knowledge, and this network is rendered partially determined through the superimposition of culturally defined rules of grouping and difference that direct the relationships between the symbolic associations, rendering the extension of some symbolic relationships possible, and others culturally unlikely. The extension and elaboration of these same symbolic association networks appears to occur in a systematic manner in the cognitive-linguistic development of small children, and it is in this process of elaboration that we can locate the theory of the symbolic differentiation of the phenomenal field. A case in point, involving the study of colors and color terms, and color-trait associations to a variety of objects and traits, illustrates a process which can be demonstrated with a variety of domains of cultural knowledge.

As the result of the analysis of a series of color related tasks which involved for the most part the use of the same values and sets of colors across the same basic sets of samples, a general theory of the symbolic use of color emerged, a theory which is largely consistently with the principles of marking and differentiation outlined in the introduction. Color is a useful means of gaining an empirical understanding of the cognitive and perceptual dimensions of symbolic and affective relationship.

The theory of color is referred to as a set of coordinate and interrelated principles of grouping and contrast between colors and sets of colors which span and differentially carve up the otherwise continuous natural spectrum of color.

Simply put, colors occur in basic dyadic sets that serve to demarcate a specific region or range of the color continuum. Some sets appear more basic and stronger than others, and some sets seem not to usually occur. Colors and dyadic sets of colors also occur in basic sets of contrasts by which different regions of the continuum become demarcated as different. Thus the conception and coordination of color demands the complex and frequently implicit interactions between basic relationships of grouping and contrast, and these interactions may be partially determined by implicit principles which may in part be naturally based and in part culturally defined.

It is by these systematic principles of grouping and contrast which appear in color associations that we can gain a handle onto the psychological significance of color as it is used, its underlying semantic dimensions. Furthermore, it allows us to make certain predictions about the nature of these relationships by which we can search for patterns. Given the paradigm of all possible dyadic relationships which occur between the 12 basic colors used in the following tasks, we can expect certain dyadic combinations to recur with strong correlational strength within correlational matrices across a wide variety of samples and tasks. At the same time we can expect certain contrastive correlations to recur as well between basic colors which appear to not occur in close association. Finally, we can expect the paucity of significant correlations in those color combinations which do not appear to be reinforced naturally or culturally and which do not seem to have a strong correlation.

The paradigm of dyadic relationships between the twelve colors above gives 66 possibilities--of these possibilities only a certain percentage (25/66 = 37.87%) might be expected to have a strong positive relationship, while the balance have either a contrastive relationship or no relationship at all, while of this subset, possibly 16 (16/66= 24.24%) might be expected to have an especially strong or naturally occurring relationship. This paradigmatic patterning of the dyadic relationships between colors function something like a list of "distinctive features" in the sense that either a relationship holds or it cannot be said to hold.

The problem is that these relationships are not truly dichotomous or obligatory, but are a question of frequency patterning and percentage of occurrence. What can be expected is that certain dyadic combinations of colors should have a repeated significant correlation by some means of statistical analysis. The resulting matrices and networks of color combination and choice define discrimination tables which are open and only partially deterministic and relatively non-hierarchical in structure.

Furthermore, we may speculate that: 

1. Some dyadic relations are more basic than others.

2. Sets of dyadic relationship can occur in "complexes" or clusters of interrelationship.

3. To some extent the strength of the dyadic relationship will determine the frequency of expected occurrence of the paired colors as paradigmatic alternates.

 4. Dyadic pairs of colors form "chains" by which different sets of colors may be indirectly interlinked. We might speculate at a higher level of analysis triadic complexes with high correlation may frequently occur between colors. 

5. A number of extrinsic variables may interfere and influence the resulting correlational matrices of these colors, such as extraneous trait associations, culturally defined dimensions of contrast, or psychologically defined patterns of saliency or symbolic significance of certain colors. 

6.The sequence of both the developmental and cultural acquisition of color terms follows the principles of contrast and similarity defined by basic inter-color associations.

These patterns of inter color association form an underlying symbolic structure with both cognitive and affective components and accretions, that appear to play a part in the framing and construction of other types of symbolic frameworks, and that become to some extent hard wired in the brain.

The preceding diagram is an attempt at a three dimensional spherical representation in which polar opposites represent basic dyads and basic contrasts between dyads are cross at right angles along an intersecting plane of the sphere. Stronger, more basic relationships are located on spheres of greater distance from the origin than weaker or more general relationships and by the heavier weight of the connecting lines.

The principles of patterning that underlie the organization and meaning of color are neither complex or of great number, nor are they obligatory. They summarize decision rule trees defined from alternate pathways of a discrimination table that describes the pattern of grouping and contrast that occurs with color. The psychological, cultural and extraneous trait associations which appear to influence this patterning of color are also not complicated, and may be described in a fairly specific manner. From these basic principles of patterning and variables it is possible that an expert system model of color can be constructed.

Certain principles may be speculated on at the outset.

1. There is a functional and interdependent relationship between the grouping of colors in basic dyadic sets, their extension in larger groupings, and the contrast between colors and sets.

2. Principles of basic grouping are relatively more independent variables than principles of contrast--i.e. dyadic grouping and its extension underlies and partially determine the contrasting of different colors.

3. Principles of color contrast are relatively dependent variables. Contrasts tend to occur between groupings of color before they occur between basic colors, and they tend to occur between non-basic colors and groupings before more basic colors or groupings

4. We may distinguish between basic and elaborated groupings of colors, as dyadic, direct associations versus triadic, indirect associations.

5. We may distinguish between neutral, nonspecific contrasts between colors, and non-neutral, specific contrasts which are direct juxtapositions of contraposed colors.

6. Non-neutral contrasts tend to involve basic sets in direct juxtaposition, and can be considered to be symbolically and projectively "loaded" and of greater significance.

7. The indirect extension of basic dyadic sets allows the substitution of paradigmatically alternate colors in contrastive color juxtapositions or constructions.

8. Non-basic, complementary relationships may occur between different colors, either grouped or contrasted, as color primes recur in more than one basic dyad become "lost" and latent in the complementary relationship.

9. The extension of these symbolic color complexes to other non-color traits is accomplished principally through the means of the quaternary structure of analogical association, i.e. apple is to banana as red is to yellow.

10. The resulting "nonverbal" language of color is structured analogically in coordinate interrelationship of colors to other colors--i.e. within a color context.

11 Symbolic associations and accretions to color therefore follow and are structured by the underlying principles of grouping and contrast in color.

12. Symbolic associations and accretions to colors, whether they are psychologically idiosyncratic or culturally shared, influence the patterning of color in fairly specific and determinable ways: i.e., the relative level of color-terminological differentiation of the color field, patterns of color preference, coordination and ranking, patterns of color-trait association, and the order, organization and dimensions of color-dimension associations.

From the analysis of the data and the pattern, certain dyadic relationships appear in a sequential order of saliency/basicness in the twelve color tasks:

1. red-yellow versus black-white

2. blue-green versus purple-violet

Grey and pink and orange and brown especially for the Chinese sample appear to be of least contrastive significance or salience, and imply a weaker dyadic relationship which falls apart more frequently. Thus these colors which are themselves non-basic appear as the least salient and become more associated with other basic colors--orange or brown with yellow or red, pink with purple or violet and gray with blue or green or with black or white.

From this schema we may infer a scale of basicness between colors and dyads. This scale appears to be for the six dyads: black/white, red/yellow, purple/violet, blue/green, pink/gray and orange/brown. In the contrasts which occur between these colors we cannot clearly separate the basic dyads, but it appears the entire set is being contrasted in the form of the quaternary structure of analogy in which basic principles of commutation and transition are being observed:

1.

red is to yellow         as blue is to green         as pink is to gray

as

white is to black     as purple is to violet     as orange is to brown

2.

red is to white as blue is to purple as pink is to orange

as

yellow is to black as green is to violet as gray is to brown

We may then cross between sets in a systematic manner:

3. red is to blue as green is to pink as gray is to red

as

white is to purple as violet is to orange as brown is to white

or, alternatively

4. blue is to white as pink is to violet as red is to brown

as

red is to purple as green is to orange as gray is to white

Such a structure can be easily elaborated if we imagine the addition of other intermediate dyadic sets, such as purple-pink, red-orange, red-violet, black-gray. It appears that culture and psychological variables intervene to influence this process of elaboration of color associations in definite ways. Though there are no hard and fast a priori rules of association, and any such a structure easily breaks down in a non-deterministic manner, we may say that in such a manner inter-color associations may be easily extended, as long as they tend not to contradict more basic contrastive sets that occurs--such as between red-yellow/black-white, and that the entire structure always tends to be anchored by more enduring sets of basic dyadic primes.

We can extend this quaternary structure to incorporate triadic sets:

red is to yellow is to orange blue is to green is to brown

as

black is to white is to gray purple is to violet is to pink

Thus:

red complements orange as blue complements brown

as

black complements gray as purple complements pink

and

orange contrasts black as brown contrasts purple

as

red contrasts gray as blue contrasts pink

Such relationships appear nonsensical from a rational standpoint, but it must be kept in mind that these associations are essentially non-rational and therefore non-logical in character. It is the non-logical, analogical character of the patterning of color that permits black apples and blue bananas to occur as valid possibilities color-trait associations though no such associations appear to occur naturally in the world.

It is at this point when we consider the influence of psychology and culture that the patterning of color really becomes interesting and problematic. It appears that though almost any possible trait-color associations can be imagined, some are more basic and anchored to reality than others, and hence serve to stabilize the structure. The association of green and brown with trees and grass are a clear and strong example for which we may imagine the following structure:

green is to foliage is to grass as yellow and orange are to sun and rays

as

brown is to trunk is to earth as blue is to clouds is to sky is to water

It is interesting that in the color-trait associations of the House-Tree-Person drawings, both foliage and trunks appear as brown, and foliage appears as green, but trunks do not appear as green. A slot is created in the sample which demands "filling" in with an explanation. We would expect that slot to also occur with high frequency, the fact that it is absent raises a fundamental question.

brown is to trunk brown is to foliage

as

green is to foliage green is not to _____???_____

Among Chinese samples, the association of black to eyes, hair and eyebrows and figure outline is strong, as well as the association of brown to skin, face, arms, hands, or pink with shirt and pants and purple with shoes:

black is to eyes is to hair is to eyebrows pink is to pants and shirt

as

brown is to skin is to face is to arms is to legs purple is to dress to shoes

The association of yellow to red permits the frequent possibility of a red sun, and the apparent association of the sky and clouds to the sun permits the possibility of a blue sun. The association of the skin to the clothes makes possible the association of pink with brown, but though we have a frequency pattern of pink faces and skin, there is no similar association of brown clothes:

brown is to face is to skin is to limbs pink is to face and limbs

as

pink is to shirt and pants. brown is not to _____???___

Again a gap has been created in the structure that we can expect to be filled in but which for some reason does not. The creation of such gaps appears to be a normal consequence of the extension of such quaternary structure, and appears to be a basic process in developmental acquisition.

It is reasonable to assume that affective associations follow similar paths as do the cognitive-semantic associations in the construction of colors and color-trait patterns, and that the symbolic construction of color association involves just such a system of transference of affective as well as semantic association of meaning from one color to another. In such a way, color is used to reinforce affectively and symbolically the transference and identity of symbols, such that:

the purple house is to the purple flower is to the purple face

as

the black figure is to the black outline of a tree is to the black outline of the house.

or

the eye is to the window

and

the curtains are to the eyebrow

as

to the center of the flower is to the sun

and

the rays of the sun are to the petals of the flower

In such a manner we can describe the symbolic construction of complementary symbolisms which have no clear, valid or logical identity but which achieve a symbolic isomorphism of identity by means of the associative transference of meaning. The fact that there are "core" underlying patterns of inter-color, inter-trait and color-trait associations confers a degree of consistency and deterministic plausibility to such structures that would otherwise be lacking.

In the preceding structure, two conjunctive "and" relationships were introduced in place of the analogical "as", and suggests the possibility of the systematic substitution of conjunctive-disjunctive connectors in place of the "as" in the increasing logical determinacy of the patterning. "As" as a connector is a more general and less obligatory relationship than either of the logical connectors. But the substitution of "and/or" relations for the "as" allow us to group and contrast sets of analogical relationships in a manner which more gives greater constraint to the resulting pattern when we substitute all "as" relationships for either an "and", "or" or a contrastive "versus" to more accurately depict the structure which actually appears to be most salient in the pile sorts of the 12 colors:

1.

red as yellow as blue as green as pink as grey

as

white as black as purple as violet as orange as brown

becomes

red and yellow or blue and green or pink or gray

versus versus or

white and black or purple and violet or orange or brown

We may refine the model a little further by introducing inferential "if-then" syntax into the model to define some deterministic rules of order of the relationships:

and and or and

(red and yellow) and (blue and green) and (orange or gray) then| (yell or gray....)

and and or and

(black and white) and (purple and violet) and (gray or brown) then| (white or brown....)

(yellow versus black) and (green versus violet) and (orange or pink) then|______?????______

The interesting aspect of such a logical structure is that it may be read simultaneously either left to right or top down, and along alternate pathways. There is no prescribed order of reading the structure, and it permits the accumulation of conclusions at the end of each sequence and in the lower right hand corner. It may be said that there are a set of logical connectors (as, versus, or, and) and if-then. It general, and is the most obligatory relationship and describes a close association. Or describes a weaker relationship between colors, and versus describes a contrastive relationship. The "as" implies an open analogical relationship which encompasses all other possibilities of relation. Parenthetic sets are nested within the larger framework, and in the reading of the thing more specific operators of relationship should take precedence before, and hence are nested within, more general operators.

Within this structure, any relationship becomes possible, as long as it is not contradicted by or contradict an initial premise of contrast. Principles of contrast are thus vital to the order and functional coherence of such a system, as it prevents the occurrence of circular contradictions which would render the structure incoherent. It is interesting that this model resembles somewhat the method of structural analysis developed by Claude Levi-Strauss for the interpretations of myths.

The structure is descriptive versus "predictive" or prescriptive in nature, but it permits the possibility of the systematic creation and the inferential filling in of gaps which may occur within the structure. Such structures quickly bog down in the tremendous exponential explosion of possibilities that they encompass. It is fortunate that such symbolic constructions tend to be neither very deep or broad. Another set of possibilities exists when we consider that the same basic inference structure may be written in several alternate ways. The fact that not all possibilities can occur with equal plausibility confers meaning and a minimal order to its patterning of relationships Such a structure also suggests the possibility of the application of non-parametric statistics in its analysis, and in the association of parallel discrimination tables that express each relationship as an estimated probability or actual frequency of occurrence, and from which decision trees and basic functional rules can be systematically deduced.

To summarize the results of these color tasks so far, it appears that it is not too difficult to imagine a green or an orange or red or blue or purple or brown or even a black banana, though it may be easier for us to recognize and eat a yellow banana. An orange may be red or blue or green, but is easier to distinguish from an apple or a ball or an unripe grapefruit if it is orange. This suggests that the semantic associations with color are quite variable and largely analogical in nature--we may imagine the possibility of black or purple apples quite easily, though there may be some imagined color-object combinations which may violate certain implicit restrictions of acceptability.

These patterns suggest that a number of principles of contrast may be at work underlying the differentiation of the color field, and the relationships between these dimensions may be neither simple nor straight forward. The following model of the dimensions of possible contrast places along a number of multi-dimensional axis those most salient sorts of contrastive variables which may influence the functional differentiation of color, and it suggests that if we are to find color in the brain, we must look in more the one place.

Names for these natural sets might include grue (green/blue), "vurplet" (violet/purple), "bronge" (brown/orange), "bred" (brown/red), "grink" (pink/grey), "blite" (black/white), "rellow" (red/yellow). Other sets may be imagined. These sets are interlocked and occur on an implicit level of organization.

The differentiation of the continuum may proceed on two levels simultaneously or in tandem--first top-down with the distinction of basic color sets, and then bottom up with the selection out of basic color from these sets. At each subsequent level, the largest complex of chains tends to be broken at the "weakest" dyadic relationship which also tends to be the mid-point separating the continuum into more basic sets or colors. It is predicted that this splitting point will be the "best subsequent discriminator"--that breakage point (or minimal points) which will result in the most coherent (least noisy) sub-groupings. Thus the color continuum "pulls itself apart" by polarization into smaller sub-groupings that come to be increasingly less implicit and more basic in terminology. The main split at each level comes to be seen in terms of the differentiation of the most salient color contrast at that level.

The only exception to this rule in the cluster analysis above appears to be the early separation of orange and yellow, of yellow from red and orange from grue. It is difficult to imagine an association between orange and green or blue.

Though the actual patterning of color association at each level of the pile sort is more complicated than this, there are basic patterns which are definitive--such as the lack of certain colors at certain subsequent levels of discrimination.

It is extremely difficult to interpret in a conclusive manner the combined results of the different tasks. Certain inter-correlations of specific dyads of colors do appear to be consistently pronounced across different samples and tasks, but are significantly influenced by variables of sample size, type, design of the task, and the association of traits to colors.

Several presuppositions behind these analysis must be brought into focus. The patterns presented in these analysis have been based upon the relative frequency patterns across samples. Individual variability has therefore been lost. There is no saying that culture is necessarily only what is shared, or is what is represented by shared frequency patterns of response to such tasks as presented above. Culture does appear to have an influence upon such patterns, but these are difficult to separate from psychological variables which also influence them. This analysis can only be considered preliminary, problematic and just scratching the surface of an extremely complex problem of analysis and interpretation.

A similar type of analogical chaining of associations was found to occur in the domain of familial terms and relationships for a Chinese sample, in relation to a number of linguistic framing tasks (grids, dichotomous, apperception, sentence completion). This chaining appears to be ordered on the basis of certain elicited rules of relationship which regulate the associations. The centrality of this kind of familial model to understanding the dynamics of Chinese culture suggest that the basic mental model is behaviorally and symbolically extended and elaborated on the organic level of institutional social patterning of a culture, and that there is dynamic feedback between the sub-organic, symbolic and organic institutional levels of information patterning within cultural systems.

There is the strong suggestion of the development of an alternative system of artificial intelligence which to some extent predicts and models the patterning of color choice and ranking above, based upon attribute-value notation, the sample frequencies and the use of a discrimination function that estimates the level of inter-color association, the probability and noise factor of each color in each position, and which searches and traces pathways through the color array, deriving rules based upon relative probabilities and likelihood of alternative pathways through the array. The same system could be set up to model and predict patterning in different tasks, and may derive rules that may otherwise remain implicit in the complex patterning of frequency distributions. It is suggested that some pathways through such discrimination networks may be more likely and better predictors than others, and many pathways are highly unlikely while a few may emerge as clearly significant.