Semiotics&kinship

Semiotic Grammar: Binary Representations of Marked and Unmarked Classes1 Paul Miers, English Dept. Towson Univ., Towson, MD 21252 [email protected]

These semantic aspects of communication are irrelevant to the engineering problem. The significant aspect is that the actual message is one selected from a set of possible messages. The system must be designed to operate for each possible selection, not just the one which will actually be chosen since this is unknown at the time of design. .

C. E. Shannon, “A Mathematical Theory of Communication”

This document presents a model for understanding the semiotics of socially constructed classification system in terms of binary trees and constraints on the representation of marked and unmarked features. I start with long standing problem from cultural anthropology, kin classification, and show how variation in a few simple constraints on a binary tree can generate very different representations of social kinship and marriage preferences. I then apply this same model to a binary classification tree built from two different underlying structures – the high order classes of the pure and the impure and a tree for marked and unmarked features of male and female classes. Finally, I apply that tree to the film Pretty Woman (1990) to show how that film begins with contention over the representation of gendered purity and concludes with a paradigm that results from moving selected marked and unmarked classes from one side of the tree to another. I suggest that this final classification paradigm reflects a social ideology where elite (pure) consumption is set against the common (impure) exchange of money for sex. The paradigm undercuts whatever claims might be made that the film disrupts the social order it reflect by first allowing a prostitute to expose the narrow minded thinking of the elites and then to join them as the pair bonded partner of their most marked male specimen.

1

Text is available under the Creative Commons Attribution-ShareAlike License. See Terms of Use for details.

2 I use the examples from kin classification because they provide concrete illustrations of certain key issues I want to discuss here: 1. Binary structures can be used to build a classification tree which has at its terminal nodes an ordered set of multiple classes. Although we may start with a conventional gender binary of male/ female, the terminal class structure of the tree can be much more complex. An asymmetrical structure such as that used in the kin classification tree can generate three terminal classes from a descending series of binary partitioning of the topmost, dominant category space. 2. Coding is determined by the dominance hierarchy of the classification tree. This concept is regularly mystified in semiotic theory whenever various “codes” are somehow pulled from a hat as needed to support a particular interpretation. Where exactly does the encoding and decoding take place? Where do the codes themselves actually live? In a classification tree the “codes” are simply the internal class structure of the tree which determines how much information is carried by a particular signifier relative to some structured set of classes. To encode is to generate terminal nodes for an output representation; to decode is to extract the information in the tree from the signifier. This narrow definition of coding, far from limiting the possible interpretations of the signifier, is what allows us to account for how a signifier can bear information from many different systems of representation The English kin term “aunt,” for example, is a lexeme which carries with it the encoded information – G+1collateral female. That information resides in the dominance hierarchy of classes created by the classification tree for English kin terms. To decode “aunt” as also being “mother’s or father’s sister” requires accessing a genealogical representation of kinship which is itself a socially constructed representation based on “folk” beliefs regarding shared substance, procreation, etc. Just as with spoken and written language, we are always dealing with what Bakhtin called heteroglossia, i.e. complex, mixed forms of representation. 3. Changing a very few constraints on a class expression can lead to radically different encodings. In some classification grammars, constraints on markedness force classes to be broken up and recomposed such that signifiers appear to “move” from one part of a tree to another and merge with members of another class. The most striking example of this move and merge phenomenon in kin classification appears in the example of “Dravidian” kin classification described here where constraints which ban marking a distinction between lineal and collateral classes causes some collaterals to become members of a parent class for ego and then causes some of ego’s “cousins” to be marked as “sibs.” This “covert” movement is directly connected to the “overt” exchange of marriageable children according to a very patterned economy, one which Levi-Strauss saw, under the rubric of the giving and taking of wives, as the fundamental social contract.

3

4. The underlying logic of class membership is controlled by the expression or neutralization of marked features. This logic, which is much more than a“primitive” analogic form of thought, lies at the core of the classification mechanism, and there is reason to believe that all such mechanisms share this logic because they evolved from a ancestral cognitive tool kit that social primates used to categorize both each other and the many “natural” and “constructed” things they consume and exchange. Cognitive scientist have focused a great deal of attention on the “prototype” problem, i.e. the problem of determining exactly what kind of mental representation counts as being the representative case or exemplar for a class. Although this problem is indeed an important one, focus on it has drawn attention away from the logic of how equivalence classes partition a domain and how selection of a marked or unmarked element from one side of a binary partition can alter the representation of all members of a higher order class. The standard notion of an equivalence class is given by:

where an equivalence relation ~ is defined on the set X such that the equivalence class of a in X is the subset of all elements in X which are equivalent to a.. The binary of the marked and the unmarked can thus be understood as governing the selection of the element a from a set because it has some feature or property which partitions the set X into disjoint subsets (a is marked as the special case ) or taking a as having some property common to all members of the set of (a is unmarked). Here I am concerned not with the question of which properties of a are common to a higher order class and which marked property partitions that class into disjoint subsets. Rather I want to focus on what happens when the lower order partitioning is blocked and yet a marked feature from one side or the other of that partitioning surfaces in the higher order set as a property common to all members of the class simply by dint of their membership in that class. Consider, for example, a class distinction I will be using here between the pure and impure as applied to the high order category of fruit: fruit

pure

impure

unmarked

marked

unmarked

generic pure fruit

special pure fruit

generic impure fruit

marked special impure fruit

4 In this case, then we have four possible terminal figures or signifiers: 1) the generic or prototypical pure fruit – e.g. a generic red apple without discoloration, etc.; and 2) a specially marked fruit which makes salient some distinctive property of purity in fruit – e.g. a bowl of perfectly ripe strawberries or bright red cherries. Likewise, we can have the generic impure or rotten fruit or some special marked case of impurity. Now these four terminal classes really give us too much information if all we are after is a signifier for impure fruit that we shouldn’t eat. Thus the three tiered tree can effectively be reduced to a two tier tree where we selectively insert or move up one or the other of the pairs of terminal nodes. Given our concern with impure fruit, the most efficient class representation would be: fruit

pureunmarked

impuremarked

generic pure fruit

special impure fruit

In other words, we want the general paradigm for pure fruit and a special case representation of some marked feature that best predicts impurity: a generic apple and a rotten banana with black splotches, for example. This sort of variable logic for forming equivalence class representations clearly has an evolutionary efficacy to it. Sometimes we might want to use the fully extended four terminal tree to get a useful sample that lets us understand some class partitioning of a cognitively salient category. In other cases, the two tiered tree is more efficient – we have an image of the generic or prototypical pure fruit and the special case of some fruit marked for a critical feature indicating we should not eat it. Consdier, however, how this variable logic of equivalence class formation can also be manipulated by constraints on the structure of the classification tree itself. For the three tiered tree above, suppose we have a constraint that bans the division between the pure and the impure. That constraint thus leaves us with a dominant top most category – fruit – and elements designated by the four terminal nodes which must collapse upwards as equal members of the top most category:

5

fruit

unmarked

marked

unmarked

generic pure fruit

special pure fruit

generic impure fruit

marked special impure fruit

In this case we have competition between the marked and the unmarked versions of the pure and the impure to be a , the element which defines the equivalency relationship on the entire class “fruit.” Suppose then that we impose another constraint: the impure must be marked so that the two constraints are ordered: no pure/impure partitioning > the impure must be marked. Satisfaction of this constraint order yields: fruit marked impure That extremely efficient one tiered classification tree might be useful if the crisis of impure fruit were so severe that the safe option was to simply avoid eating fruit. But it can also be used to induce such behavior even if there is no impurity crisis. Given the constraint order, all elements in the set Xfruit are presumed by their very membership in that now unpartitioned set to be impure. Moreover, the property or feature that makes fruit impure need not be discernable or even testable. It may be some unspecified genetic mutation or herbicide; or it simply may be a curse or interdiction on all fruit. It does not matter. We have a particularly nefarious kind of reification: all fruit is now marked impure by the equivalency relationship of being a member of the class “fruit.” Consider what happens when this same logic is applied to humans with respect to the contested issues of ethnicity, gender, and social class. If we block or ban the unmarked generic representation, then we can always define a class by way of some marked feature. If the special case is the socially valorized half of the binary, then we can effectively indict all other members of the class for failing to be the special case. If the special case is drawn from the devalued half of the binary, then we effectively move all members of that higher order class into the devalued partition even though the partitioning that divides some category into some valued and devalued classes is no longer part of the coding. .

6

a) genealogical tree generation+1

generation0

Δ

Ο

Ο

Δ

Δ

Ο

MB

MZ

M

F

FB

FZ

Ο

Δ

Ο

Δ

Ο

MBD

MBS

MZD

MZS

Z

ego

Δ

Ο

Δ

Ο

Δ

B

FBD

FBS

FZD

FZS

M = mother; MZ = mother’s sister; MB = mother’s brother; F=father; FZ = father’s sister; FB = father’s brother Z = sister; B= brother; Δ = male; Ο = female MBD = mother’s brother’s daughter; MBS = mother’s brothers son; MZD = mother’s sister’s daughter; MZS=mother’s sister’s son FBD = father’s brother’s daughter; FBS=father’s brother’s son; FZD=father’s sister’s daughter; FZS = father’s sister’s son

b) kin classification tree

generation+1

rank

▲ lineal+1

collateral+1

“parent”

“parent’s sibs”

lineal0

♀

collateral0

♂

= lineal0 sex

♀ ego

♂

= lineal+1 sex

≠ lineal0 sex

♀

♂

♀

♂

♂

cross & parallel sibs” order

♀

≠ lineal+1 sex

cross & parallel cousins ►

Fig. 1 – genealogy and kin classification The top figure (a) shows the standard two generation genealogical tree anthropologists have used to represent kin classification systems which assign kin terms (e.g. uncle / aunt) to kin types (mother’s brother / mother’s sister). What has long been puzzling is the fact that classification systems vary widely in how they map kin terms to kin types and, more importantly, many of them appear to ignore or violate genealogically defined classes. The bottom tree (b) is a binary classification tree which can capture the logic of the genealogical tree and yet can generate different classification systems depending on how constraints are defined on the use of lexically interpretable classes in the tree, that is on classes which reflect information that is directly encoded in the kin term.

7

generation+1

a) lineal paradigm

lineal+1

collateral+1

lineal0

collateral0

ego

sibs

cousins

b) G+1 kin terms

generation+1 lineal+1

collateral+1

♀ mother

♂ father

♀

♂

aunt

uncle

generation0

c) G0 kin terms

lineal +1

collateral+1 cousin

lineal0 ♀ V ♂ ego

collateral0

♀

♂

sister

brother marry out

►

Fig. 2 – lineal kinship – no parity classes; single collateral class composed of “aunts and uncles;” neutralizes both sidedness (patrilateral / matrilateral ) and consanguinity / affinity (by blood / by marriage) within collaterals. English kinship terms; variations also found in some hunter-gatherer bands.

8

generation+1

lineal+1 lineal0

collateral+1

collateral0

ego

sibs

= lineal+1 sex

≠ lineal+1 sex

♀+1

♂+1

♀+1

♂+1

paternal parallel cousins

maternal parallel cousins

paternal cross cousins

maternal cross cousins

generation+1 lineal+1

collateral+1

♀ ana

= lineal+1 sex

♂ baba

≠ lineal+1sex

♀+1

♂+1

♀+1

♂+1

diaza

apça

hala

dayí

generation0

lineal+1

collateral+1

lineal0 ♀ V ♂ ego

lineal+1 =sex

collateral0

elder

younger kardaş

♀ aga

♀+1

♂+1

diaza çocuğu

apça çocuğu

▲

▲

≠ lineal+1 sex

♀+1 hala çocuğu

♂+1 dayí çocuğu

▲

♂ aba marriage by collateral lineage

Fig. 3 – bifurcate collateral - maximize collateral classes; no lineal parity classes; creates possible marriages with one of four collateral lineages. Common in Mideast, also China; example here is Turkish.

▲

9

a) bifurcate merging kin classification generation+1

≠ lineal+1 sex parents’ cross sibs

= lineal+1 sex

parents & parallel sibs

= lineal0 sex

≠ lineal0 sex same sex cross cousins

cross sibs

ego & parallel sibs

opposite sex cross cousins

b) G+1 kin terms

generation+1 = lineal+1 sex

≠ lineal+1 sex

♀

♂

iten

timin

c) G+2 kin terms

♀

♂

uhun

un

generation 0 = lineal+1 sex

= lineal0 sex

elder noatun

younger noatahan

≠ lineal+1 sex

= lineal0 sex

≠ lineal0 sex

≠ lineal0 sex

♀ ego

♂ ego

♀ ego

♂ ego

♀ ego

♂ ego

nauvnen

namanin

newum

nevin

rahniaruman

rahnpetan

▲

▲

can marry can marry

Fig. 4 - bifurcate merging – no lineal / collateral class division; parents are merged with same sex sibs; ego belongs to a class with same sex sibs; merges same sex parallel cousins into that class; creates extended marriage exchanges which blend lineages and often generations; example here is from Telugu – a language spoken in south India.

10

human

♂

♀

unmarked

marked

unmarked

marked

generic male

special male

generic female

special female

purity

pure

♂

impure

♀

♂

♀

unmarked

marked

unmarked

marked

unmarked

marked

unmarked

marked

generic pure guy

special pure guy

generic pure girl

special pure girl

generic impure guy

special impure guy

generic impure girl

special impure girl

Fig. 5 – Gendered Purity. The two tier “human” category with marked and unmarked options for male and female nodes generates four terminal types. When this tree is joined to a simple one tier tree for the pure/impure binary, eight terminal types are generated.

11

purity pure

impure

unmarked

♀ marked

♂

all males → generic pure guy

all females →special impure girl

a) simple two class binary tree with all males unmarked for purity; all females must be marked for impurity

purity pure

♀

impure

marked

♂unmarked

all females →special pure girl

all males → generic impure guy

b) simple two terminal class binary tree where all females must be marked for purity; all males unmarked for impurity

purity pure

impure

♀ unmarked ♂unmarked all males → generic pure guy

♀marked some females → special pure girl

most females → generic impure girl

c) binary tree with three terminal classes: males unmarked for purity; pure females must be marked; unmarked females are impure.

Fig. 6 - Possible grammars for marked / unmarked human purity

12

Pretty Woman (1990) synopsis from Internet Movie Database; film available on YouTube playlist http://www.youtube.com/view_play_list?p=945F716DA38262F2 A very successful, wealthy lawyer, Edward Lewis, hires a beautiful and unlikely prostitute, Vivian Ward (Julia Roberts), from Sunset Blvd to bring along to various business events. An attraction develops between the two, and Edward finds it harder and harder to let the infectious, kind-hearted Vivian go. Businessman Edward Lewis (Richard Gere) breaks up with his girlfriend, who doesn't want to be at his "beck and call" at a swanky party held by his partner Philip Stuckey, and cuts loose in Stuckey's Lotus. He gets lost and stops along Hollywood Boulevard for directions from hooker Vivenne (Julia Roberts). She charges, gets in and ends up driving him to his hotel in Beverley Hills. He asks her up to his penthouse suite on a whim and pays her to stay all night, although seems uncomfortable at first. The following morning he asks her to stay all week for $3000. He also gives her money for clothes and says she needs to be at his "beck and call" with no strings attached. She calls her room-mate Kit (Laura San Giacomo) to leave her money for the rent, and goes shopping on Rodeo Drive for more appropriate clothes. However, snooty saleswomen won't serve her as she is still dressed like a hooker, and she returns to the hotel, where she gets stopped by the Hotel Manager, Barney (Hector Elizondo). He wants to make it clear that they are making an exception having her at the hotel as Edward is such a special guest. Vivienne gets upset as she still has no outfit for dinner, and Barney helps her, along with coaching her on dinner etiquette. When Edwards returns, he is amazed by Vivienne's new look. The business dinner goes well, but Edward is preoccupied with the deal afterwards. The next day, Vivienne tells him about the experience shopping the previous day, and Edward takes her back to spend an obscene amount of money on clothes, leaving her to go back to his work as she is transformed from hooker to lady. She goes back the shop from the previous day to show them the big mistake they made! Back at hotel, she looks like a genuine guest, but when Edward gets home he is still busy with work, and they take a bath together and talk into the night about their pasts and how they ended up where they are today. The following day, Edward takes Vivienne to the polo. While Vivienne chats to David Morse, the grandson of the man involved in Edward's latest deal, Philip is worried she is a spy. Edward reassures him by telling him how they met, and Philip then comes on to Vivienne. When they return to the hotel, she is furious with Edward for telling him, and plans to leave, but he persuades her to see out the week. Edward leaves work early the next day and takes Vivienne on a date to the Opera in his private jet. She clearly is moved by the music, and says "If I forget to tell you later, I had a wonderful time tonight". On returning to the hotel, he falls asleep (the first time we have seen this) while she is getting ready for bed, and she kisses him on the lips - she doesn't do this with clients - and they make love as partners, rather than client and hooker. Over breakfast, Edward offers to put her up in an apartment so he can continue seeing her, but she feels insulted and says this is not the fairytale she wants. He then goes off to work without resolving the situation. Kit comes to the hotel and sees that she has fallen for him, but she denies it. Edward meets Morse, about to close the deal, and changes his mind at the last minute. His time with Vivienne has shown him another way of being - taking time off and enjoying life - and working. He wants to create things rather than just making money. Philip is livid, and goes to the hotel. Vivienne is there and he blames her for changing Edward - he comes onto her again, and then hits her before Edward returns and pulls him off and chucks him out. Vivienne leaves, and is seen back at home with Kit, packing up to leave for San Francisco. Edward gets into the car

13 with the chauffeur that took her home, and rather than going to the airport, he goes to her apartment and climbs up the fire escape (despite being afraid of heights) with a rose in his mouth, to woo her like in a fairy-story.

purity

pure

♂

impure

♀

♂

♀

unmarked

marked

unmarked

marketed

unmarked

marked

unmarked

marked

generic pure guy

special pure guy

generic pure girl

special pure girl

generic Impure guy

special impure guy

generic impure girl

special impure girl

Philip

Edward

Edward’s girlfriend

?

drive by johns

Carlosthe dealer

Kit

Vivienne

Fig. 7 - Starting classification grammar for Pretty Woman Pure world – marked and unmarked males compete for unmarked females Impure world – marked males dominate marked and unmarked females All marked females are impure; some marked males are pure; some unmarked females are impure purity

pure

♂ marked

Ø

impure

♀marked

♂ unmarked

♀ unmarked

special pure guy

special pure girl

generic impure guy

generic impure girl

Edward

Vivienne ▲

Philip ▲

Kit

Fig. 8- Concluding classification grammar of Pretty Woman Pure world – marked males pair with marked females. Impure world – unmarked males pair with unmarked females. Purity is always marked; impurity is never marked

Ø

14

The marked impure girl who gives her clients colored condoms. She must be checked at the clinic once a month.

Purification of the marked impure girl. The bubble bath jouissance of becoming Julia Roberts

15 Briefly, the account of the Pretty Woman goes as follows. The pure / impure binary in this film partitions a world into two domains, one defined by purity of taste for elite commodities (cars, penthouses, champagne and strawberries, etc.) and the other by the impurity of sex for money which requires protection from disease. We start with the typical male and female buddy pairings: two males inhabiting the world of pure, high end products and values, and the two prostitutes living off the streets in contention with a drug dealer. When the film begins both of these worlds are overpopulated, but there is one figuration missing from the pure world: the marked female. All women in the pure world are unmarked as generic members of a social class. Edward’s unmarked girl friend leaves at the very start of the film and the implication is that all unmarked generic women in this world will marry someone else. It is not so much that Edward resists domestication but that none of them are good enough for him to break through the barrier of his secretary. This “full” representation can be taken as the input, and the final representation is the output which satisfies the constraints or grammar of this world. That grammar turns out to be rather simple: purity is always marked; impurity is never marked. Thus not only does Edward’s generic girl friend leave the scene, but Phillip, the generic sidekick, is shown to be the very embodiment of the common impure male. The one marked male in the impure world, Carlos the drug dealer, is neutralized by the gun carried by Edward’s black chauffeur. And finally, of course, the marked impure girl becomes the marked pure girl because Edward returns to climb the fire escape and fetch her from a life in the impure world. Phillip and Kit are presumably left behind in the impure world as the common female for whom sex is work and the common male who must work to get sex. The political ideology of this input / output transformation seems evident enough. The pure world is a restricted domain filled with rare individuals who act out the fantasy of high end consumption which the impure world of the commoners can never directly experience except by way of a fantasy: the jouissance of being Julia Roberts as the object of Richard Gere’s desire. The grammar of Pretty Woman, that is, just is that of Hollywood itself. Moreover, these rare elites, our elites, when allowed to run the order of things, have conveniently neutralized the one disturbing element which might make the necessary continuation of impure world dangerous for the stars: the marked impure male.