THE TROBRIAND CASE

Cultural Inferences

 

by Hugh M. Lewis

University of Missouri, Columbia

1995

 


The Trobriand Ethnographic Context

 

Edwin Hutchin's work Culture and Inference (1980) deals centrally with a land claims case in a traditional open moot setting in the Trobriand Islands. Rights to individual garden plots are divided into the right to use and the right to allocate the garden. While the right to use the garden may be transferred back and forth between dalas, or clans, the right to allocate a garden must be kept within the same dala and can be lost only to the village headman upon the death of the owner. Such a complex set of rights and rules of transference of gardens yields a complicated history of the "pathway" any particular garden plot takes in the course of several generations.

There have been important ethnographic precedents in understanding this Trobriand system of land rights--most notably B. Malinowski (1935) and Annette Weiner (1976). Previous accounts failed to evaluate the functioning of the system in an adequate manner because they did not analytically separate the rights to use and to allocate garden plots.

There are, as well, other facets of Trobriand and wider Melanesian culture which need to be taken into account in regard to the cultural appropriateness of the cases to be considered. These considerations of context broach an important aspect of the study, of its possible cultural relativity as something uniquely Trobriand or typically Melanesian, immersed inextricably in a holistic web of other traits and relations. It brings up the plausible generality of the model, and of the somewhat paradoxical question of "how much context" is enough in our translations, interpretations, models and theoretical representations of any cultural logic.

The central theoretical issue is, of course, about the underlying and largely implicit logical coherence of culturally encoded schemata that allow for the public consensus and individual integration of experience and manipulation of the common stock of knowledge.

"The analysis of litigation has shown that a model of folk logic developed from purely western sources is quite adequate as an account of the spontaneous reasoning of Trobriand Islanders. It is not straight Aristotelian logic, because it contains plausible as well as strong inferences, but then so does our own reasoning. There is no need to posit a different logic....The clear difference between cultures with respect to reasoning is in the representation of the world which is thought about than in the processes employed in doing the thinking. It is clear that Trobrianders cut the world into a different set of categories from those we entertain, and that those categories are linked together in unfamiliar structures. But the same types of logical relations underlie the connections, and the inferences that are apparent in their reasoning appear to be the same as the inferences we make." (Hutchins, 1980:127-8)


Basic Metaphors

Hutchins bases his interpretation of this system upon the operation of certain logical connectives which work like key inference eliciting metaphors and which are used in a prototypical manner at several levels of implication.

The metaphor of keda or the pathway of the garden, is an important one in understanding the relevant schemata and basic model presented in the text. This metaphor of the pathway is reiterated in several places in the text, and is seen as an example of an even more powerful abbreviating mechanism of cultural codings than schematic chunks because it is able to compress whole episodes into a single statement. "The keda metaphor allows the speaker to refer to the individual episodes in the discourse in terms of the most salient events within them." Determination of the keda, or history of the social pathways a garden will make through the generations, constitutes the principle basis for the litigation of disputes over land-claims. The success of a litigant often depends upon his/her ability to reconstruct a credible history of transactions--"such a history will typically trace the keda (path of social movement) of the garden only a few generations back to some previous rights holder whose claim is acknowledged by both litigants." (Hutchins, 1980:45)

Malinowski (1935) noted the frequency of the theme of quarreling over gardens in Trobriand folklore. This metaphor may be important, not just for understanding a broader Trobriand cultural network of exchange--when we speak, for instance of the pathway of a particular armband or shell necklace in the far-flung Kula Ring--but for understanding of more general process of cognitive mapping in an individualís effective life-world, as for instance, the multiple, interconnected pathway a young child learns to take in her/his adaptive navigation of a myriad of different, new experiences.

Another important metaphor is that of pokala and pokala exchange forms the primary schemata of the cultural model. It was Motabasiís misinterpretation of Monilobuís gift as "pokala" that constituted part of the basis for his defeat in his claim. There is a linguistic confusion between semantic levels of meaning of the term, between a general, unmarked, basic form, and a specific, marked, explicit form. "In the most general sense of the word, pokala denotes any prestation from an individual of inferior status to one of superior status in the hope, but without the promise, that something will be returned."(Hutchins,1980:25-6)

The general, unmarked form of pokala serves as the basis of several constrast sets formed in relation to other marked kinds of exchanges such as katuyumali or katumamata. Hutchins presents a diagram (1980: 28) that illustrates pokala as an "umbrella term" which as an unmarked member is at the center of the intersection of three dimensions: 1) institutional context of exchange; 2) the medium of initial exchange; 3) relation of participants to land. These dimensions are the basis of different contrast sets in wider domains of exchange, one whose terms distinguishes between various expected returns, another whose different terms distinguishes the various media of exchange, and the third whose terms distinguish the social loci of use rights to the garden plot. (1980: 27) Different pokala arrangements serve as alternative "roads" or pathways to the acquisition of rights to property that "are not traveled with equal frequency"(1980: 43)

The standard pokala is at the heart of interpersonal politics and thus of the basic generalized reciprocities of social relations. It serves as the prototypic form of social transaction "between a man and his motherís brother or his motherís motherís brother." It is the most frequent form of exchange for land transfer that occurs--approximately 80%--within the "owning" dala.

Annette Weiner remarks in her work on Trobriand culture that "The most important part of making pokala, I was told, is not to say the word pokala aloud when giving things to someone. To make pokala, a man gives and then he waits. To talk about pokala is a shameful thing to do and can often detract from success."(1976: 157)

"Pokala relations between the holders of resources and persons who aspire to those resources are not legal contractual arrangements. To see them only as legal obligations is to miss both the emotional texture of the relationships, which is itself a major determinant of the volume of tokens exchanged, and the freedom of choice available to rights holders and aspirants in the formation and maintenance of such relations. Every exchange event is a communication from one person to another of both an artifact (item exchanged) and a social message. The movement of artifacts makes exchange important economically. The participantís interpretation of social messages makes exchange important symbolically. (Hutchins, 1980:36-7)

It becomes apparent that the notion of pokala, and propositional schemata it presupposes, enters into a broader arena of social exchange, relation, systems of prestation, status, prestige and value by which Trobriand culture elaborates itself as something distinctive in the world. But this should not preclude its more general relation or relevance to other and larger cultural and symbolic systems of exchange, status and value by which people use cultural codes to construct and then make sense of their world.

The final metaphor which deserves mention is tupwa which signifies that a garden is "extra, spare, left-over" (Hutchins, 1980:36) with respect to the holder and has not yet been allocated to someone else. "As long as a garden is tupwa, it is accessible to all types of pokala."(1980:36) It was Kwaiwaiís opinion that the garden Kolubuwa was not tupwa in respect to Motabasi, an opinion which he reiterated four times in the course of his presentation, both in the opening, intermediate and closing statements, that served to clarify, as a powerful recursive metaphor, the real status and history of the garden and the insubstantial basis of Motabasiís claim to it. That the garden was not tupwa in respect to Motabasi was repeated in the Chiefís final decision in the case. At this point, the deeper cultural significance of tupwa might only be guessed at without the further elucidation of the broader ethnographic context.

Expert Systems

The effort to fit such a complex as Trobriand land claim cases into an expert system depends upon a clear and explicit model of the system formulated in terms of basic "if-then" rules and confidence factors which are logically connected. An expert system "shell" (Benfer et al., 1991) is basically a system built on the basis of the elicited knowledge and implicit rules of a human expert upon a delimited domain of knowledge, which is then emptied of its informational content and contains only its skeletal structure of an inference engine and interface.

The shell and design of the system employed was "backward chaining." A backward chaining system is one that is "goal driven" in the sense that it works from one of a few alternative goals, taken as the terminal factive consequents of a string of related "if-then" rules, and then searches backward among the number of alternative antecedents until it achieves a "factive" confirmation. The resulting search string or inference chain, starting from a limited number of input facts and a given goal, to a final conclusion linking the goal to the facts, may be quite long and convoluted in its logical permutations. The logic upon which such a system may be classical or two value, fuzzy or Bayesian.

One virtue of the interface of a backward chaining system is that if it lacks facts to connect to goals or subgoals, in solving its problem, it can be made to query the user for more inputs which it then incorporates as "facts" in the search-solution space.

There has been some debate over the question of whether such expert systems represent anything more than sophisticated decision-tree models. While a decision-tree is basically a static device and represents but one pathway through a "search solution space," a rule based expert system, founded as it is upon logical operators, is much more dynamic and represents alternative pathways from which different decision-trees may be constructed. Furthermore, the rules upon which such expert systems are built are largely implicit to the knowledge of the informant, and thus are normally out of awareness in everyday decision-making, though nonetheless determinative in their outcomes. Decision-tree models based upon conscious decision-making models simply do not represent this implicit level of structural analysis. Furthermore, expert systems may be "generative" in that they may lead to the formulation of new rules and the acquisition of new facts which might otherwise be oblivious to the informant.

The important aspect of expert systems design in regard to their application in ethnographic research is their role as mediational devices which allow the concise and coherent translation of models or constructs of understanding between a user and a foreign knowledge domain, or alternatively, an explicit knowledge of the ethnographer and the informant's mostly implicit cultural knowledge. If such systems can be made to work, they can be said to constitute a kind of test of the coherence or non-contradiction of the underlying logic of the knowledge structure. Whether or not this test of the model also constitutes a test of the empirical validity or ethnographic consistency of the theory on which it is based, is another, possibly unanswerable matter.

The Basic Model

The basic model as elaborated in chapter three of Edwin Hutchins' work Culture and Inference (1981:46-61) concerns the acquisition, transfer and retention of rights to the use and allocation of garden land by members of different dalas, or clan groups in the Trobriand Islands. Rights to use such plots of land are separate from the rights to allocate such land, and are subject to a corporate, trans-generational history of title of the land as well as to a record of the ritual exchanges which have been made, or claimed to have been made, in reference to either the use or allocation of the land.

This history of land use, transfer, and entitlement, and the distinction drawn between the right to use (R-use) and the right to allocate (R-allocate) serves as the point of contention in competing claims to the land, as well as the basis for the settlement of such disputes by an open moot officiated by tribal officials. The record of rights of entitlement to land is in a sense a matter of public knowledge, and is thus subject to the common cultural constraints, sanctions and evaluations.

The model consists of a set of primary and secondary derivative schemas which serve as the foundation for deciding the probability of possession and transfer of these rights of allocation and use. The basic "master" schema involves the case of member B of dala one giving "pokala" to member A of the same dala in return for the allocation of both rights to allocate and to use the garden and can be written in shorthand as:

 

A (R-allocate + R-use)

B gives pokala to A

B receives (R-allocate + R-use)

 

This schema provides the basis for a set of strong inferences to be made regarding the likelihood of certain events, either the actual transfer of rights to B, as well as the previous conditions which constrain the transfer--that A originally has both rights to use and allocate the property and that B gave pokala to A. If either A lacks rights to allocate or use the garden, or B does not give pokala to A, then it is certain that A could not in the first case, and would not in the second case, transfer rights to B. These corollaries can be written as:

 

1. A (No R-Use or No R-Allocate) & 2. A(R-use +R-allocate)

B gives pokala to A B does not give pokala to

B does not receive rights B does not receive rights

 

Similarly, it can be assumed that if B has rights to the land or is using the land that was previously possessed by A, then at some previous time B had given pokala to A, and A transferred all rights to B which A had previously held.

The first set of derivative schemata involve the distinction between transfer of rights within a dala, which are transferred intact, versus the transfer of only use rights between dala, and which rights are therefore split between rights of allocation kept within the dala and rights of use given to the other dala. Rights of allocation cannot be transferred outside of a dala. This can be modeled as follows:

 

2. A[dala one](R-allocate + R-use) & 1. A[dala one] (R-allocate + R-use)

B[dala one] gives pokala to A B [dala two] gives pokala A

A transfers (R-allocate + R-use) to B. A transfers (R-use) to B.

 

In the second case above, the rights to a garden plot become split between two dala. In the case of a third party (C) giving pokala to either A of dala one or B of dala two, the resulting transfer will be only of either rights to allocate, in the first case, and only the rights to use in the second case, but not both.

 

1. A[dala one] (R-allocate) & 2. B[dala two] (R-use)

C[dala one] gives pokala to A C[dala two] gives pokala to B

A transfers (R-allocate) to C B transfers (R-use) to C

 

Several counterfactuals can be inferred to be true from these schemata. In the first case, C will not receive the right to use the garden, and, in the second case, C will not receive the right to allocate the garden.

"Katumamata" involves the "waking up of a previous pokala" by a third party of the non-owning dala (C [dala two]) to the owning dala, and involves the subsequent transfer of the the rights to use the garden from B to C.

 

C[dala two] gives katumamata to A[dala one]

B [dala two] transfers (R-use) to C

C receives (R-use) from B.

 

"Katuyumali" involves the transfer of the use rights back from the non-owning dala to the owning dala. This schema is represented thus:

 

B [dala two] (R-use)

A[dala one] gives katuyumali to B [dala two]

B transfers (R-use) to A

 

The final schemata involves the death of the owner of the land and the appropriation of the rights to the land by the headman.

 

A[dala one] {or B or C} (R-allocate + R-use)

A{or B or C} dies

(R-allocate + R-use) transferred to Headman.

Computer Representations of the Model

It is at first appearance a relatively simple and straight-forward matter to transcribe this basic model directly into terms appropriate for an expert system shell such as Intelligent Developer, (or, alternatively and preferably, Prolog).

Each of the basic schemata (pokala, katuyumali, katumumata) can be represented by a basic rule set which covers not only the schema itself but all implicit inferential presumptions which underlie and constrain the truth-value of the schemata, such as the estimates of the likelihood of a transfer or possession, or the likelihood of certain past events like the giving of pokala or katumamata given current uses or claims.

Finally, the schemata, when linked together in some kind of order, or "history," which each garden possesses, yields a simple decision tree structure which represents the possible pathways by which the movement of the rights to land can be traced from one party to another. The number of possible pathways followed by the land rights, in the ideal case, becomes rapidly increased with the number of transfers and instances of ritual prestation, and the number of participants or the time depth involved, and resembles the basic "search tree" that is so important in artificial intelligence design.

In the case of mapping this basic model, certain initial presumptions have been made to simplify the resulting tree:

 

1. The prototypical model is written in terms of a single piece of land, and thus the schematic history of this single property is represented by one search tree.

2. "A" is the original owner of the property, and is a member of dala 1 which originally held rights to both allocate and use the garden.

3. No one has died.

4. Katumamata or katuyumali is given only to possessors of rights to the garden.

5. "C" will always be of the same dala, or clan, as "B". If "B" is of dala one, then so will be "C" and if "B" is of dala two, then so will be "C".

 

These presumptions keep the model simple--if any of these presumptions are voided, then the resulting possible set of pathways becomes increasingly complicated, and the number of rules required to adequately represent the resulting possibilities rapidly increases to the point of being unwieldy by single goal oriented expert system, and a multiple goal system would have to be constructed.

The main goal of the design is to ascertain: 1) who has what rights, if any, such that all peopleís status in relation to the land are stated and no two people share the same sets of rights, as fitting the paradigm of table one; and 2) four primary pathways of transfer of rights to the garden can be followed, as fits diagram two below.

 

Table I

Paradigm of Possibilities

 

Persons

R-intact

R-use

R-allocate

no-rights

deceased

A

yes/no

yes/no

yes/no

yes/no

yes/no

B

yes/no

yes/no

yes/no

yes/no

yes/no

C

yes/no

yes/no

yes/no

yes/no

yes/no

Headman

yes/no

yes/no

yes/no

yes/no

N/A.

From Table One above it is readily apparent that there are many possible combinations of who has what rights, with each combination representing a separate pathway to be reckoned. The primary pathways represented in Figure One below show only four alternative set of combinations of rights as illustrated by the different outputs (Output A, B, C, and D).

It is apparent that many more combinations, such as introducing third or fourth members who are making claims to the land, would produce a very complicated and difficult to read map.

There is also a necessary ordering of modus ponens type rule sets ( if-then conditional clauses) which any backward chaining expert system must follow such that certain basic "facts" must be adduced before others can be reasonably concluded. This sets in motion a built-in constraint to the scale and scope of an expert system as a knowledge based system working within a well-defined domain of inputs and rules, such that there is an a finite number of rule sets and pathways which can be optimally handled within a single goal paradigm. Too many alternative conclusions, entailing multiple sub-goals and numerous alternative pathways, especially when these occur in the first order rules connected directly to the given facts or when alternate goals may share the same rules, rapidly reduces the representational capacity and resolving power of the model as an adequate construction of reality.

We have soon reach the familiar Von Neumann bottleneck of a resulting combinatorial explosion of the search space, without any clearly obvious or "correct" functional algorithm to resolve this dilemma. This exploding complexity would be reflected in the number of non-contradictory rules necessary to control the system, and in the number of the conditional antecedents and concluding consequents of each rule.

The following is a cursory outline of the rule sets, and definitions provided for Intelligent Developer that is illustrated by table one.

BEGIN. [Start]

Output. (Welcome) Pause. Output (Initial Presumptions) Pause. Output(Pathways). Pause.

1. A is of Dala one.

2. No one is deceased.

3. If B is of dala one, then C is of dala one.

4. If B is of dala two, then C is of dala two.

5. Only B makes transaction to A, and C makes transaction only to A.

 

1. Did B give pokala to A? Pokala Yes; pokala no.

a. rule: If B gave pokala to A, then transfer of rights was likely.

b. rule: If B did not give pokala to A, then there was probably no transfer, and A retains rights intact. Output I. Pause. End.

2.Is B of dala one or dala two? Dala one; dala two.

c. rule: If B is of dala one, then transfer of rights is intact.

Output(Indefinite). Pause. Seek(three).

d. rule: If B is of dala two, then transfer of rights to use only. Output(Indefinite). Pause. Seek(four)

3. Did C give pokala to B? Pokala yes, pokala no.

e. rule: if C gave pokala to B, then transfer of rights intact. C holds rights intact. Output II. Pause. End. f rule: if C did not give pokala, then no transfer of rights. B holds rights intact. Output III. Pause. End.

4. Did C give katuyumali to B? Katuyumali yes, katuyumali no.

g. rule: if C gave katuyumali to B, then C transfer rights use only. Output(Indefinite). Pause. Seek (six).

h. rule: if C did not give katuyumali to B, then no transfer of rights. Output(Indefinite). Pause. Seek (five).

5. Did A katumamata B? Katumamata Yes, Katumamata No.

i. rule: if A gave katumamata to B,then B transfer rights use back to A. A retains rights to allocate and receives rights to hold, A holds rights intact. Output IV. Pause. End.

j. rule: if A did not give katumamata to B, then B retains rights to use the garden and A retains rights to allocate the garden. Output V. Pause. End.

6. Did A katumamata C? Katumamata Yes, Katumamata No.

k. rule: if A gave katumamata to C, then C transfers rights to use the garden back to A. A retains rights to allocate and receives rights to hold. A holds rights intact. Output. VI Pause. End.

l. rule: if A did not give katumamata to B, then C retains rights to use the garden and A retains rights to allocate the garden. Output.VII. Pause. End.

Complete. There is not enough information for I.D. to decide who has what rights. Would you like to choose? A_intact; B_intact; C_intact; B_use, A_allocate; C_use, A_allocate.

Endrule. If [END] , then Output(Finished). Pause.

Output I. Because there was no pokala given, there was no transfer of rights, and A retains rights intact to the garden. Format(1). (A_intact, B_none, C_none)

Output II. Because C gave pokala to B who received rights intact from A, chances are that C was granted rights intact to the garden. Format (2) (C_intact, B_none, A_none)

Output III. Because C did not give pokala to B, C receives no rights to the garden and B holds rights intact. Format (3) (B_intact, C_none, A_none)

Output IV. Because A gave katumamata to B, B probably transfered use rights back to A, and A therefore holds both rights to use and allocate the garden. Format (4). (A_intact, C_none, B_none).

Output V. Because A did not give katumamata to B, B retains rights to use the garden and A retains rights to allocate the garden. Format (5) (A_allocate, B_use, C_none).

Output VI. Because A gave katumamata to C, A receive back the right to use the garden while retaining the right to allocate the garden. A therefore holds rights intact to the garden. Format (6) (A_intact, B_none, C_none).

Output VII. Because A did not give katumamata to C, A does not receive the right to use the garden, which is retained by C. A has only the right to allocate the garden. Format (7) (A_allocate, B_none, C_use).

Output (Indefinite) It is yet uncertain who has what rights. More information must be found.

Given the preceding outline, the expert system design follows the following Figure Two of the possible pathways in the system:

Inference Values and Confidence Factors

An important aspect of this entire analysis is the use of inference in cultural schemata--the use of a notation (s, p, T, and F) prefixed to each clause denoting the degree of inference strength associated with the statement--strong, plausible, strongly True, strongly False, plausibly True and plausibly False, True and False. The possible inferences ranging from weakest to strongest are:

strongly false (sF); plausibly false (pF); plausible (p);

plausibly true (pT); strongly true (sT).

One means of representing these inference values in Intelligent Developer are the use of confidence values which can be arbitrarily assigned on a percentage basis to each of the conditionals and conclusions of each rule.

strongly F. =10%; plausibly F. = 30%; plausible =50%;

plausibly T. =70% strongly T. =90%.

It is these uncertainty factors which drive inference in both the cultural system and in its computer representation. These uncertainty factors are rooted in the lack of public knowledge of the actual history associated with a garden, and in alternate claimants manipulation of this knowledge in establishing the credibility of their own interests. Given that a pokala has been made by a member of the same dala, the transfer of complete rights intact is almost obligatory, and thus receives a rating of strongly True; but if no pokala has been made the claims of a recipient's rights to a garden will be regarded as plausibly False. Cases of cross-dala transfer of rights, or back-transfer following katuyumali, are less certain and entail plausible inference values; transfer of use rights in these instances are not as obligatory as within dala transfers.

 

Test Cases

These basic rule sets, if designed correctly and properly transcribed in the form of Intelligent Developer, should hypothetically serve as an "inference engine" by which to test the efficacy of the model as it has been applied in the case analysis provided in chapter 4 of Hutchin's book which presents a running dialogue of a land claims case. If the basic models of the expert systemís are correctly constructed according to the logic of the basic schema, it should be possible to predictively test the logical chain of each of the following presentations to see if the expert system will reach the same set of conclusions. In the case of Intelligent Developer, this also requires that the exact ordering of choices be predetermined for each presentation, based upon the rule ordering of the program.

Claimant's argue for rights to use and/or allocate the garden based upon different interpretations of the history of previous possession, prestations and transfers associated with the garden. The interpretation of the history of the garden, because publicly unknown, can be manipulated to support different claims to the garden.

The presentation by the alternate claimants in the case are divided into episodes by which they build their claims and establish public credibility to rights to the garden.

Motabasiís Presentation

The first presentation is by Motabasi who is of the same dala as the original owner but who had not previously used the land and who had left the land untouched for many years but who then returns to reclaim and reawaken past rights to the land.

Episode One

Motabasi claims that Motolala, Woiyaii and Taubagoni's rights to allocate and use the garden are false, and that Motabasi holds these rights.

But Motabasiís claim is in question. The previous possessor of garden is uncertain, as well as the Brothersí claims to have given pokala in return for transfer of rights:

?{A + U(garden)}

Brothers give pokala to ?

? transfers to Brothers {A + U(garden)}

Episode Two

Motabasi establishes previous possession of garden:

 

Older Brother{A+U(garden)}

Ilawokuva plausibly gives pokala to Older Brother

P[Older Brother transfers rights to Ilawokuva]

P[Ilawokuva{A+U(garden)}

 

Episode Three

Motabasi establishes his claim to garden:

 

Ilawokuva{A+U(garden)}

Motabasi gives pokala and kaivatam to Ilawokuva,

by helping the old woman garden the land.

Ilawokuva transfers rights to Motabasi

Motabasi{A+U(garden)}

 

Episode Four

Motabasi interprets a prestation by Monilobu of a different dala to Ilawokuva to be pokala, and not katumamata, and therefore to be plausibly false:

 

Ilawokuva{A+U(garden)}

Monilobu gives pokala to Ilawokuva

Ilawokuva plausibly transfers rights to use the garden to Monilobu

p. F[Monilobu{U(garden)}

 

Episode Five

Motabasi claims that he gave sufficient pokala to Ilawokuva who transferred full rights to him:

Ilawokuva{A+U(garden)}

Motabasi gives pokala to Ilawokuva

Ilawokuva transfers rights to Motabasi

Motabasi{A+U(garden)}

 

Episode Six

Motabasi strongly disclaims his younger brotherís receiving rights from their Sister because they did not support her:

Ilawokuva{A+U(garden)}

F(Brothers gave pokala to Ilawokuva)

s. F(Ilawokuva transfers rights to Brothers)

p. F(Brothers{A+U(garden)}

 

Episode Seven

Motabasi constructs a counterfactual schema in order to later refute it in episode eight:

 

If(Inaveguwa{A+U(garden)}

and Brothers gave pokala to Inaveguwa

Then p. T(Inaveguwa transfered rights to Brothers) and

p. T(Brothers{A+U(garden)}

 

Episode Eight

 

Motabasi asserts that it was Ilawokuva who held rights to the garden, and implicitly not Inaveguwa, therefore his claims are valid and the counterclaims are false:

 

Ilawokuva{A+U(garden)}

Motabasi gave pokala to Ilawokuva

Ilawokuva transfers rights to Motabasi

T(Motabasi{A+U(garden)}

 

Episode Nine

 

The schema in episode nine is implicit in the refutation of the counterfactual construction of episode seven by the reassertion of the claims in Episode Eight:

 

F(Inaveguwa{A+U(garden)}

if(Brothers give pokala to Inaveguwa)

then s. F(Inaveguwa could have transfered rights to Brothers)

and p. F(Brothers{A+U(garden)}

 

Motabasi's hypothetical history of the garden can be summarized in the following Figure Three:

Kailimilaís Presentation

Kailimila is the counter-claimant to the case and has received a tactical advantage over Motabasi in getting to present his claims after Motabasi. Kailimila has a certain credibility given a public history of his use of the garden and the long term absence of Motabasi.

 

Episode One

Kailimila opens his presentation by ridiculing Motabasiís assertion that the people he listed whose claims to the land are false:

F(F(Motolala, Woiyaii, Taubagoni, Inaveguwa{A+U(garden)})

Kailimila then reports an earlier and important conversation between himself and Ilawokuva while Motabasi was still living away in another village, in which she instructs him to recover the garden from Solubwaís group who were of a different dala and who gave pokala to her for use rights to the garden. By this report Kailimila establishes that Ilawokuva was the previous possessor of the rights to the garden, and by which is instantiated an entire set of schemas that establishes his claim to the garden:

Ilawokuva{A+U(garden)}

Solubuwa, of a different dala, gives pokala to Ilawokuva

Ilawokuva transfers rights to use the garden to Solubuwa

Solubuwa{U(garden)}

 

Episode Two

Ilawokuva{A(garden)}

Kailimila gave pokala to Ilawokuva

Ilawokuva transferred rights to allocate garden to Kailimila

Kailimila{A(garden)}

 

Episode Three

 

Kailimila{A(garden)}

Ilawokuva{U(garden)}

Kailimila gave katuyumali to Solubuwa

Solubuwa transfers rights to use the garden back to Kailimila

Kailimila{A+U(garden)}

 

Episode Four

Kailimila reconstructs a hypothetical instruction to him by Ilawokuva which would have made matters easier for him, involving an earlier katumamata given by Monilobu to receive use rights to the garden:

 

Ilawokuva{A(garden)} and Solubuwa{U(garden)}

Monilobu gave katumamata to Ilawokuva

Solobuwa transfers rights to use the garden to Monilobu

Monilobu{U(garden)}

 

Episode Five

Kailimila reports the proceedings between himself and Monilobu, when he was instructed by the court to give katuyumali to Monilobu to recover rights to use the garden:

Ilawokuva{A+U(garden)}

Solubuwa gave pokala to Ilawokuva

Ilawokuva transferred rights to use the garden to Solubuwa

Ilawokuva{A(garden)} and Solubuwa{U(garden)}

Kailmila{A(garden)} and Monilobu{U(garden)}

Kailimila gives katuyumali to Monilobu

s. T(Monilobu transfers rights to use the garden to Kailimila)

s. T(Kailmila{A+U(garden)}

 

Episode Six

Kailimila recounts the actual public event of giving katuyumali to Monilobu and his giving up rights to the garden, and then turns to attack Motabasiís claims by comparing these to his own previous public record of rights to the garden. Kailimila ridicules Motabasiís claim to the garden by the weakness of his assertions of never having gardened it and never publicly receiving rights to it:

Kailimila{A(garden)} and Monilobu{U(garden)}

Kailimila gives katuyumali to Monilobu

Monilobu transfers rights to use the garden back to Kailimila

s. T(Kailimila{A+U(garden)}

 

Episode Seven

Kailimila recounts the schema for within dala transfer of rights to the garden(kuluboku) and that the particular garden in question was not among those allocated to Motabasi by Ilawokuva:

 

Ilawokuva{A+U(Wa, Kap, Bwei, [but not Kb] )}

Motabasi gives pokala and kaivatam to Ilawokuva

Ilawokuva transfers rights to Wa, Kap, and Bwe, (but not Kuluboku) to

Motabasi

Motabasi{A+U(Wa, Kap, Bwe)}

 

Episode Eight

Kailimila recounts again the public katuyumali he gave to Motabasi in exchange for use rights to the garden, thereby conclusive reassrting his main claim to hold rights intact to the garden.

Kailimila{A(garden)} and Monilobu{U(garden)}

Kailimila gives katuyumali to Monilobu

Monilobu transfers rights to use the garden back to Kailimila

s. T(Kailimila{A+U(garden)}

 

The following Figure Four represents the reconstruction of Kailimila's hypothetical history of the garden:

Insert Figure IV here.

 

Kwaiwaiís Final Opinion

Kwaiwai gives the final reconstruction of the most plausible history of the garden, thereby deciding the case in favor of Kailimila and against Motabasi's claims.

 

Episode One

Kwaiwai gives an initial hypothetical construction representing Motabasiís presentation:

If Ilawokuva{A+U(g)}

and Motabasi gave pokala and kaivatam to Ilawokuva

Then p. T (Ilawokuva transferred rights to Motabasi)

p. T(Motabasi{A+U(g)}

 

Episode Two

Kwaiwai states that it is false that Ilawokuva did not have rights to the garden to transfer them to Motabasi, therefore Motabasi did not come into possession of the garden following his pokala:

F(Ilawokuva{A+U(g)})

Motabasi give pokala and kaivatam to Ilawokuva

p. F(Ilawokuva transferred rights to Motabasi) and

p. F(Motabasi{A+U(garden)}

 

Episode Three

Kwaiwai sets up a model to compete against the hypothetical construction of episode one:

Ilawokuva{A+U(garden)} but

? someone had already transferred the rights to the garden.

 

Episode Four

 

Kwaiwai reiterates the hypothetical construction of episode one to emphasis its hypothetical basis:

If Ilawokuva{A+U(garden)}

and Motabasi had given pokala and kaivatam to Ilawokuva

Then p. T (Ilawokuva would transfer rights to Motabasi) and

p. T(Motabasi{A+U(garden)}

 

Episode five

Kwaiwai begins to elaborate who had already transferred the rights to the garden:

Ilawokuva(dala one){A+U(garden)}

Solubuwa(dala two) gave pokala to Ilawokuva

Ilawokuva transfers right to use the garden to Solobuwa

Ilawokuva{A(garden)} and Solobuwa{U(garden)}

 

Episode Six

Kwaiwai confirms Motabasiís argument that one arm of bananas is insufficient as pokala for the garden, making it plausible that Ilawokuva did not transfer rights to use the garden to Monilobu:

Ilawokuva{A+U(garden)}

Monilobu gave pokala(one arm of bananas) to Ilawokuva

s. F(Ilawokuva transfers rights to use the garden to Monilobu

p. F(Monilobu{U(garden)}

 

Episode Seven

 

Kwaiwai reinterprets this schema as katumamata rather than the mistaken presumption of Motabasi that it was pokala:

Ilawokuva{A(garden)} and Solobuwa{U(garden)}

Monilobu gave katumamata or tiatavan to Ilawokuva

p. T(Solobuwa transferred rights to use the garden to Monilobu)

p. T(Monilobu{U(garden)}

 

Episode Eight

Kwaiwai uses the metaphor of "keda" to refer to the pathway of the garden, and thus, the author claims, carries the whole discourse organization to a higher level that involves recursion among the propositions themselves.

 

Episode Nine

Kwaiwai reiterates the hypothetical construction of episode one:

If (Ilawokuva{A+U(garden)}) and

Motabasi gave pokala and kaivatam to Ilawokuva

Then p. T(Ilawokuva transfered rights to Motabasi)

p. T(Motabasi{A+U(garden)}) and

p. F(Kailimila{A+U)garden)}

 

Episode Ten

Kwaiwai then reiterates the schema of episode five, in which the rights to use the garden had already been transferred to another dala:

Ilawokuva{A+U(garden)}

Solobuwa(dala two) gives pokala to Ilawokuva

Ilawokuva transferred rights to use the garden to Solobuwa

Ilawokuva{A(garden)} and Solobuwa{U(garden)}

 

Episode Eleven

 

Kwaiwaiís statement refers to the whole episode five of Kailimilaís presentation, in which Kaimila describes his recovery of the garden from Monilobu:

Kailimila{A(garden)} and Monilobu{U(garden)}

Kailimila gave katuyumali to Monilobu
Monilobu transfered right to use the garden back to Kaimila

s. T(Kailimila{A+U(garden)})

 

Episode Twelve

 

Kwaiwai finally and conclusively reiterates the initial hypothetical construction:

If Ilawokuva{A+U(garden)} and

Motabasi gave pokala and kaivatam to Ilawokuva

then p. T Ilawokuva transferred rights to Motabasi

and p. T(Motabasi{A+U(garden)} ) and p. F (Kailimila{A+U(garden)})

 

Kwaiwaiís presentation represents a succinct, authoritative and comprehensive summary of the history of a garden, one which resembles almost exactly Kailimilaís summary, except that the former incorporates the conclusive schema of Motabasiís relationship in the affair. The hypothetical case of Motabasiís rights it presents is necessary to the understanding the basis for the trial and the conflict of interests between competing claimants to the garden, and for showing how this case did not fit the history of the path of the garden.

Kwaiwai's opinion is presented in the following Figure Five showing the accepted history of the garden:

Insert Figure V here.

Intelligent Developer can be used to test Motabasiís presentation because it is the shortest one, and basically involves only three participants. If the following paradigm of responses correctly provides for the questions that the expert system asks, the expert system should conclude the same claim that Motabasi concludes with.

 

1. Assume that Ilawokuva is (A), that Monilobu is (B) and that Motabasi is (C), and that (A) and (C) are of dala one, but (B) is of dala two.

2. (B) gave pokala(insufficient) to (A), but (A) did not transfer rights to use the garden to (B).[In the program, B did not give pokala to A]

3. (C) gave pokala to (A), and (A) transfered rights intact to (C).

4. (C) has rights intact to the garden. [(B) has no rights to garden]

 

Given this scenario, the computer should conclude with C receiving rights intact from A, and B having no rights to the garden.

The following rule sets, presuppositions and responses represent the attempt to apply Intelligent Developer to this summary history and need to be given in order for Intelligent Developer to adduce the correct decision:

1. Ilawokuva, Kailimila and Motabasi are of dala one: Solubuwa and Monilobu are of dala two.

2. Ilawokuva is A(dala one); Kailimila is B(dala one); Motabasi is C(dala one); Solubuwa is B(dala two): Monilobu is C(dala two).

3. Solubuwa gives pokala to Ilawokuva, who in turn transfers Use rights; Monilobu gives katumamata to Ilawokuva, and Solubuwa transfers Use rights to Monilobu; Kailimila gives pokala to Ilawokuva, and Ilawokuva transfers rights to allocate the garden to Kailimila; Motabasi gives pokala to Ilawokuva, Ilawokuva holds no rights to the garden, and therefore Motabasi receives no rights. Kailimila gives katuyumali to Monilobu.

4. A special rule needs to be added to account for Motabasiís false claim. This rule is to be written as "If C(dala one) gives pokala to A and A holds no rights to garden (because of previous pokala) then C receives no rights to the garden).

 

General Considerations of Theory and Method

There may be no one correct way to construct an expert system of a representation, or a representation of a model, or a model of a text, or a text of a translation, or a translation of another text, or a text of a dialogue, or a dialogue of an event, or even an event itself, much less an expert system model of a complex cultural event. Somewhere in the long chain of inference and implication is a wide latitude for alternative variation and recombination. As we move from the concrete realities of phenomenologically rooted experience toward increasingly abstract representations of that experience, we risk the greater possibility of error, bias, extraneous influence, deviation, over-reduction, reification, arbitrariness, and metaphorical and evaluative looseness.

With increasing simplicity and sophistication we risk greater spuriousness, superficiality and unrecoverable loss of the sense of complexity inherent in the original experience or event. And yet, in spite of all our constructions there is a need to maintain a sense of convergence which is never satisfactorily exact but always heuristically useful. We can preserve the "essential" elements and do without the clutter in our most abstract representations, and still have confidence that they say something significant, if not quite profound or substantive, about our common reality.

Edwin Hutchins makes several important points in his brief book--points which are subtle and yet profound in their implications. The logic of the basic schemata which he elucidates, if properly reconstructed, seems to work when correctly fit within the format of an expert system shell. The models prove self-sufficient , and do not seem to require any other, extraneous presumptions or information in order to make them "fit the facts." Though this tends to corroborate his conclusions, it alone does not conclusively validate his contentions of the logical order of the Trobriand system of land rights.

But concentrating exclusively upon this level of logical analysis of the cases and schematic model that he presents would be to miss entirely the most important interpretive dimensions of his contribution to the understanding of culture and cognition--the significant role which culturally constrained inference, linguistically encoded implication, and cognitively constructed "chunks" of relational schemata may play in the corporate organization of our shared experience--in the construction, instantiation, and evaluation of our basic cultural models and their derivative elaborations of belief and behavior.

In this regard, there are several points reiterated and emphasized in his work which constitute its main theoretical points.

Schematic "Chunking"

In chapter five (1980:110-124), Hutchins talks about the possible uses of the "cultural code" or "grammar" and the "chunking" of our experience of the world into culturally appropriate schemata--"a set of propositions which encode an episode" (1980: 61) that can be propositionally justified and experientially instantiated and which can be used to justify and exemplify our cultural encoded propositions about experience. "The knowledge structure is assumed to be a cognitive structure which is shared by the participants...Such a structure performs different cognitive functions depending on the requirement of the task." (1980: 70)

Such schemata "are the knowledge structures which specify the organization of culturally meaningful event sequences." (1980: 84) "The schemata provide a specification of the organization of relations among the component propositions. An ungrammatical organization lacks sensibility: it is not culturally meaningful. That is not to say that a sensible organization is necessarily true..." (1980: 61)

Such chunks are culturally prototypical and serve to linguistically and representationally abbreviate and reconstruct a great deal of information within a small, efficient cognitive space. Cultural chunking is useful in a number of ways: 1) our understanding, explanation and evaluation of events and experiences which may be problematic or ambiguous; 2) in problem-solving; 3) in our judgments of sensibility, truth-value, likelihood and plausibility; 4) in decision-making; 5) in framing our expectations and plans of future experience; 6) in our interpretation of violations of such frames of expectation; 7) in our attribution and reference of personality traits, states of being, correctness, to others and ourselves; 8) and in the interpretation of various speech acts and other symbolic representations about the world. "In the absence of a structure which specifies logical conjunctions among semantic relations in the world, it is impossible to plan, and impossible to attribute plans to others." (Hutchins,1980: 124)

Inference

The principle function of culturally encoded, schematic chunks are that they allow users to make many complex inferences about our many different experiences or events in the world with only partial background knowledge of those events, history, or hermeneutical preunderstandings and which preclude and precondition more definitive and complete knowledge. The propositional structure of such schemata allow for inference, or the association of relative truth-value and significance to both the representations and the things they represent. "Inference is the process of determining the truth value of a proposition by bringing to bear on its possible truth values the constraints imposed by its relation to other propositions whose truth values have been previously determined. In some cases the relations are such that the truth value of one proposition is completely constrained by the truth values of other propositions..." (1980: 55)

Such schemata provide systematic, ordered "plausibility structures" by which we relate to, evaluate and make sense of our experience and the events of the world. The important function of our schemata are that they allow us to make predictive, correct inferences about the world with a high rate of success.

Public Knowledge

Similar to our scientific objectivity, the most important aspects about cultural schemata and the events they stand for are that they instantiate and are based upon actual experiences which are more-or-less publicly observable. In large measure, the final decision of the cases presented rested upon the common knowledge of events observed to happen in the past. "While it is possible to use such a network of logical relations to infer the truth value of a proposition from an inferred truth value of another, the chain of inference must ultimately be anchored by a proposition whose truth value is established by observation, or lacking that, by social convention." (Hutchins, 1980: 57)

Some propositions are more easily assigned truth value on the basis of observation than others, and in cases lacking observable characteristics, truth value must be established through inference alone. "Because of the differences in observability of the propositions, some of the inferences available in the schema are made more frequently and more naturally than others..." (Hutchins 1980: 58)

Recursion

In the last part of the case analysis, Hutchinís makes reference to the meta-linguistic principle of "recursion" which refers to a "higher level of discourse organization" between propositional relations and between the schemata which encode these relations. Recursion is one of the fundamental properties of schematic propositional representations in that "since they stand for specific entities and events in the domain of discourse, they are themselves concepts and may be linked by relations to other concepts. This allows propositions to nest within each other." (Hutchins,1980: 51)

He mentions key metaphorical statements demonstrating "the heuristic power of a process of recursive embedding of progressively more complex units. The metaphor of the movement of gardens in this case allows for the concise representation and expression of relations among relations among relations among concepts." (1980: 103)

"The reader will recall that as a formal structure, the model of semantic information representation has a property called recursion. That is, a proposition relates a group of concepts to each other, and a proposition is also itself a concept. A relation among propositions is therefore also a proposition....Applying the recursion once more, we generate a proposition which asserts a relation among units, each of which asserts a relation among propositions. Each episode in discourse, being modeled on a schema for a transfer of rights, asserts a relation among a set of propositions. The structure generated here, then, asserts a relation among units of discourse...."(1980: 102-3)

Implication

Implicit throughout the work, but nowhere explicit, is the notion that implication is somehow the speakers' referential complement of the function of inference. Cultural schemata are configured from, and refer implicitly back to, a background of shared knowledge, pre-understandings, common experiences and histories. In the different presentations, the speakers are relying upon the listeners to draw the necessary, appropriate implications from the inferences and statements they are making. Indeed, in the use of counterfactuals and hypothetical constructs, inference itself is made implicit, and demands implication. In this regard, Hutchins makes reference to "scripts."

"In addition to relations within the schema, the individual propositions of the schema have relations to other propositions outside. Some of these relations might be thought of as scripts for the internal structure of the propositions. Thus, for example, there is script-like knowledge about the typical actions of a man who has use rights in a piece of land. A holder of use rights usually either gardens the plot himself or has someone garden it for him each time the major field in which it is located is gardened by the community. On the strength of his script-like knowledge, the fact that a man has not gardened a plot of land himself nor had someone garden it for him may be taken as circumstantial evidence against an assertion of his having use rights in the garden.(Hutchins,1980: 57-8)

A paradox of Hutchinís approach is that he presents an alternative, highly suggestive, and even seminal theoretical framework without focally naming it--perhaps because it is difficult to get conceptually hold of, or perhaps because it is without its own name, it seems elusive in the imagination of its possibilities. As "cultural code" he makes reference to a kind of world knowledge or common sense which is conventionally constrained in important ways. But as a "cultural grammar" it is something more than merely convention-bound knowledge--its inference generating capacity and its propositional, even logical structure, give to it more coherence as well as flexibility as a sense-making mechanism than is generally associated with plain common sense.

"Just as grammatical sentences may represent propositions which are false, meaningful arguments may come to false conclusions. A compelling deception is one that makes perfect sense. It is a common and sometimes dangerous old strategy to treat sensibility as an indication of truthfulness. Such a heuristic is understandable, though, in light of the comparative ease of judging sensibility versus judging truth in most discourse. The importance of schemata such as these is that they are standards by which sensibility is judged." (Hutchins,1980: 61)

Conclusion: The Natural Analysis of Cultural Systems.

The logical consistency of the schematic model and the test cases appear to be born out by their application in artificial intelligence programs. In terms of its logic, while one "bug" existed in the program in extending down one pathway that could not be accounted for in terms of the shell's syntax, the model otherwise appears to be complete and not to be lacking in any critical feature that would be necessary to its successful operation in a computer program.

Expert systems have been the 1980's success story in the application of Artificial Intelligence. Though the adequacy and falsifiability of such computer-based knowledge systems in really representing the actual cultural phenomena has been questioned, there has been a gradual extension of expert systems designs into ethnographic research, particularly in the modeling and rendering explicit of the expert knowledge of informants which is often regarded as ordinary and implicit.

The strength of such applications comes from the paradox that though the claim of the expert system's rules may be unfalsifiable, which if ignored, may have dangerous consequences, nevertheless they "are precisely the kinds of mediational models that ethnographers strive to formulate." (Pfaffenberger, 1988:76). The danger is to confuse and reify what is in fact only an alternative computer model of ethnographic phenomena with the "validity" of a theory that accounts for that phenomena.

But we must seek the sufficiency of the model at another level of analysis, and ask again whether or not something might have been lost in the translation and subsequent abstraction entailed by its schematic representation. Hutchins notes that translators, as analysts, have the occasion to superimpose their own logical frames upon native discourse "not so much in the translation of language per se as in whatever we do to cause our readers to generate a conceptual structure that they can use to make sense of the translation. If the ethnographer errs in the presentation of the principles that underlie the discourse, the reported discourse may still map nicely onto those principles, but its meaning--the understanding it evokes in the reader--will be different from the understanding evoked in the native by the original." (Hutchins, 1980: 48)

In the first section of his presentation of the basic cultural model, Hutchins presents a set of Trobriand logical connectives with their general English glosses and syntactic constructions. He notes the extent to which pragmatic considerations, as well as semantic understandings, are required in the interpretation of the meaning of connectives in natural discourse. "The point of all this is two-fold. First, if we are to understand natural discourse we will need to know what the logical relations are that these connectives of discourse point to. Second, if we are to understand how discourse is understood, we will have to have a way to explicitly represent these logical relations." (1980: 50)

Hutchins does not believe that there can be any direct translation of underlying logical connectives, because logic provides an appropriate meta-language "for the organization of the structures to which natural language points, but not for natural language itself. In fact, it seems that deciding on a set of direct translations of logical connectives is one way to produce a Ďwantoní translation that can make the natives seem queer." (1980: 48)

It is not too much to describe a naturalistic translation and analysis of such discourse as an attempt to represent abstractly a dynamic cultural system which is critically constrained both by conventional and contextual associations, by reference and the ground of implication this entails, on the one hand, and, on the other hand, a relational system of "logical connectives" which aims not so much at reference, as at the kind of "plausibility structures" afforded by its variable inferential functions--by considerations and evaluations that aim at the internal validity and truth value, or credibility of the statement independent of its actual empirical referents.

The former kind of meaning is situated and dependent upon its contextuality. It is always "local" in significance; the latter, like mathematical understanding, is in a sense independent of context in its fundamental "relationality"--it carries significances which tend to be "general" in character. But before we commit ourselves to an unbridgeable analytic/synthetic dichotomy, we must see that: 1) both reference and inference functions are mutually constraining and constitute a principal dialectic of discourse by which meaning is constructed and evaluated, and 2) master metaphors, or "tropes," functioning on Hutchinsí principle of recursion, such as "keda" or "pokala" or "tupwa", act as bridging mechanisms between the analytic and synthetic components of cultural constrained, cultural constructed meaning, on a level which is metalinguistic and metalogical in structure.

It is worth noting that not only are the inference strength, metaphorical salience or referential significance associated with particular schemata or connectives variable in the extent to which they are culturally constrained, but the direction of inference itself--whether right or left or upward or downward--may be determined by culturally defined syntactic-semantic considerations (Hamill, 1990.) In either case, clear ethnographic instances exist demonstrating this important linkage in the "Worldview Problem."

It is suggested that such key metaphors bridge the empirical and rational, the analytic and synthetic, the referential and inferential, respectively, by effectively combining both sets of functions, and, by virtue of their symbolic constitution and stylistic use, their symbolic function is also primarily pragmatic, versus semantic or syntactic (but not exclusively so), in its illocutionary force. It is in terms of natural language, and the metaphorical structure of natural discourse, that concepts, schema and propositions, to use Hutchinsí terminological distinctions, and reference and inference, salience and significance, meaning and value, come together into an inextricable entanglement we call understanding.

In natural language, we cannot clearly separate the one function or dimension from the other. The best we can do is to recognize and perhaps point out the prototypicality of some kinds of words and meanings as distinguished from other kinds. Again, to reiterate Hutchins, "Just as grammatical sentences may represent propositions that are false, meaningful arguments may come to false conclusions. A compelling deception is one that makes perfect sense." (Hutchins, 1980: 61)

Hutchinsí proffers a prescription for a naturalistic translation, as well as for a naturalistic ethnography of how "people reason in the real world"--by setting about to make rigorously explicit ethnographic knowledge that would otherwise remain mostly implicit only, and to construct a model which will adequately perform the functions required of the cultural knowledge.

 

BEGIN. [Start]

Output. (Welcome) Pause. Output(Initial Presumptions) Pause. Output(Pathways. Pause. 1. A is of Dala one.

2. No one is deceased.

3. If B is of dala one, then C is of dala one.

4. If B is of dala two, then C is of dala two.

5. Only B makes transaction to A, and C makes transaction only to A.

 

1. Did B give pokala to A? Pokala Yes; pokala no.

a. rule: If B gave pokala to A, then transfer of rights was likely. b. rule: If B did not give pokala to A, then there was probably no transfer, and A retains rights intact. Output I. Pause. End.

2.Is B of dala one or dala two? Dala one; dala two.

c. rule: If B is of dala one, then transfer of rights is intact.

Output(Indefinite). Pause. Seek(three).

d. rule: If B is of dala two, then transfer of rights to use only. Output(Indefinite). Pause. Seek(four)

3. Did C give pokala to B? Pokala yes, pokala no.

e. rule: if C gave pokala to B, then transfer of rights intact. C holds rights intact. Output II. Pause. End. f rule: if C did not give pokala, then no transfer of rights. B holds rights intact. Output III. Pause. End.

4. Did C give katuyumali to B? Katuyumali yes, katuyumali no.

g. rule: if C gave katuyumali to B, then C transfer rights use only. Output(Indefinite). Pause. Seek (six).

h. rule: if C did not give katuyumali to B, then no transfer of rights. Output Indefinite. Pause. Seek (five).

5. Did A katumamata B? Katumamata Yes, Katumamata No.

i. rule: if A gave katumamata to B,then B transfer rights use back to A. A retains rights to allocate and receives rights to hold, A holds rights intact. Output IV. Pause. End.

j. rule: if A did not give katumamata to B, then B retains rights to use the garden and A retains rights to allocate the garden. Output V. Pause. End.

6. Did A katumamata C? Katumamata Yes, Katumamata No.

k. rule: if A gave katumamata to C, then C transfers rights to use the garden back to A. A retains rights to allocate and receives rights to hold. A holds rights intact. Output.VI Pause. End.

l. rule: if A did not give katumamata to B, then C retains rights to use the garden and A retains rights to allocate the garden. Output.VII. Pause. End.

Complete. There is not enough information for I.D. to decide who

has what rights. Would you like to choose?

A_intact; B_intact; C_intact; B_use, A_allocate; C_use, A_allocate.

Endrule. If [END] , then Output(Finished). Pause.

Output I. Because there was no pokala given, there was no transfer of rights, and A retains rights intact to the garden. Format(1). (A_intact, B_none, C_none)

Output II. Because C gave pokala to B who received rights intact from A, chances are that C was granted rights intact to the garden. Format (2) (C_intact, B_none, A_none)

Output III. Because C did not give pokala to B, C receives no rights to the garden and B holds rights intact. Format (3) (B_intact, C_none, A_none)

Output IV. Because A gave katumamata to B, B probably transfered use rights back to A, and A therefore holds both rights to use and allocate the garden. Format (4). (A_intact, C_none, B_none).

Output V. Because A did not give katumamata to B, B retains rights to use the garden and A retains rights to allocate the garden. Format (5) (A_allocate, B_use, C_none).

Output VI. Because A gave katumamata to C, A receive back the right to use the garden while retaining the right to allocate the garden. A therefore holds rights intact to the garden. Format (6) (A_intact, B_none, C_none).

Output VII. Because A did not give katumamata to C, A does not receive the right to use the garden, which is retained by C. A has only the right to allocate the garden. Format (7) (A_allocate, B_none, C_use).

Output(Indefinite) It is yet uncertain who has what rights. More information must be found.

 

Inference values and confidence factors

An important aspect of this entire analysis has been the use of inference in cultural schemata. Important in this inference has been the use of a notation (s, p, T, and F) prefixed to each clause denoting the degree of inference strength associated with the statement--strong, plausible, strongly True, strongly False, plausibly True and plausibly False, True and False. The possible inferences ranging from weakest to strongest are:

 

strongly false, plausibly false, plausible, plausibly true, strongly true

 

One means of representing these inference values in Intelligent Developer are the use of confidence values which can be arbitrarily assigned on a percentage basis to each of the conditionals and conclusions of each rule.

 

strongly F. =10%; plausibly F. = 30%; plausible =50%; plausibly T. =70% strongly T. =90%;

 

In terms of Prolog, these inference values can be written as basic predicates, introduced as separate conditional clauses, and then written as part of the goal. It might take the following form:

 

domains

inference = symbol

 

predicates

confidence_values(inference)

holds_rights_to(person, garden, inference)

receives_rights_from(person, person, inference)

holds_no_rights_to(person, garden, inference)

 

goal

what_is_the__confidence_value_of and

write( "It is", Symbol, "that X(Y, Z)" ) and nl.

 

clauses

confidence_value(strongly_false).

confidence_value(plausibly_false).

confidence_value(plausible).

confidence_value(plausibly_true).

confidence_value(strongly_true).

 

holds_rights_to(motabasi, kolubuwa, confidence_value).

....etc., etc.

 

It should be noted that adding such inference factors considerably lengthens the number of clauses, as well as the number of rules needed to account for all the possible combinations in Prolog.

 

/* Basic Model */

 

domains

person = symbol

rights = symbol

clan =symbol

inference = symbol

 

predicates

name(person)

is_member_of(person, clan)

holds_(person, rights) receives_rights_to_allocate_from(person, person)

receives_right_to_use_from(person, person)

receives_rights_intact_from(person, person)

receives_no_rights_from(person, person)

gives_pokala_to(person, person)

gives_katumamata_to(person, person)

gives_katuyumali_to(person, person)

gives_no_pokala_to(person, person)

gives_no_katuyumali_to(person, person)

gives_no_katumamata_to(person, person)

confidence_values(inference)

 

goal

who_has_rights_to_the_garden(Symbol) and what is the confidence_value of (Symbol) and write ("It is " confidence_value "that" name "holds" rights "to the garden").

 

clauses

name(A).

name(B).

name(C).

 

is_member_of(name, dala_one).

is_member_of(name, dala_two).

 

holds(name, right_to_allocate, confidence_value).

holds(name, right_to_use, confidence_value).

holds(name, rights_intact, confidence_value).

holds(name, no_rights, confidence_value).

holds(name, rights, confidence_value).

 

rights(intact).

rights(allocate).

rights(use).

rights(none).

 

confidence_value(strongly_false).

confidence_value(plausibly_false).

confidence_value(plausible).

confidence_value(plausibly_true).

confidence_value(strongly_true).

 

gives_pokala_to(name, name).

gives_katumamata_to(name, name).

gives_katuyumali_to(name, name).

 

gives_no_pokala_to(name, name).

gives_no_katumamata_to(name, name).

gives_no_katuyumali_to(name, name).

 

receives_right_to_use_from(name, name, confidence_value).

receives_right_to_allocate_from(name, name, confidence_value).

receives_rights_intact_from(name, name, confidence_value).

receives_no_rights_from(name, name, confidence_value).

 

receives_rights_from(B, A, strongly_true) if

gives_pokala_to (B, A).

holds(B, rights, plausibly_true) if

receives_rights_from(B, A, strongly_true).

receives_no_rights_from(B, A, strongly_true) if

gives_no_pokala (B, A).

 

holds(B, no_rights, strongly_true) if

receives_no_rights_from(B, A, strongly_true)

 

receives_rights_intact_from(B, A, strongly_true) if

gives_pokala_to(B, A) and

is_member_of (B, dala_one).

 

holds(B, rights_intact, plausibly_true) if

receives_rights_intact_from (B, A, strongly_true).

 

receives_right_to_use_from(B, A, plausibly_true) if

gives_pokala_to (B, A) and

is_member_of (B, dala_two).

 

holds(B, right_to_use, plausible) if

receives_right_to_use_from(B, A, plausibly_true).

 

receives_rights_intact_from(C, B, strongly_true) if

gives_pokala_to (B, A) and

is_member_of (B, dala_one) and

gives_pokala_to (C, B).

 

holds(C, right_intact, strongly_true) if

receives_right_intact_from(C, B, strongly_true).

 

receives_no_rights_from(C, B, strongly_true) if

gives_pokala_to (B, A) and

is_member_of (B, dala_one) and

gives_no_pokala_to (C, B).

 

holds(C, no_rights, strongly_true) if

receives_no_rights_from(C, B, strongly_true).

 

receives_rights_to_use_from(C, B, strongly_true) if

gives_pokala_to (B, A) and

is_member_of (B, dala_two) and

gives_katuyumali_to (C, B).

 

holds(C, right_to_use, strongly_true) if

receives_right_to_use_from(C, B, strongly_ture).

 

receives_no_rights_from (C, B, strongly_true) if

gives_pokala_to (B,A) and

is_member_of (B, dala_two)

gives_no_katuyumali_to (C, B).

 

holds(C, no_rights, strongly_true) if

receives_no_rights_from (C, B, confidence_value).

 

receives_right_to_use_from (A, B, strongly_true) if

gives_pokala_to (B ,A) and

is _member_of (B, dala_two) and

gives_no_katuyumali_to (C, B) and

gives_katumamata_to (A, B).

 

holds(A, rights_intact, strongly_true) if

receives_right_to_use_from(A, B, strongly_true).

 

receives_no_rights_from (A, B, strongly_true) if

gives_pokala_to (B, A) and

is_member_of (B,dala_two) and

gives_no_katuyumali_to (C,B) and

gives_no_katumamata_to (A,B).

 

holds(B, right_to_use, strongly_true) and holds(A, right_to_allocate, strongly_true) if

receives_no_rights_from(A, B, confidence_value).

 

receives_right_to_use_from (A , C, strongly_true) if

gives_pokala_to (B, A) and

is_member_of (B, dala_two) and

gives_katuyumali_to (C, B) and

gives_katumamata_to (A,C).

 

holds(A, rights_intact, strongly_true) if

receives_right_to_use_from(A, C, strongly_true).

 

receives_no_rights_from (A, C, strongly_ture) if

gives_pokala_to (B,A) and

is_member_of (B, dala_two) and

gives_katuyumali_to (C, B) and

gives_no_katumamata_to (A,C).

 

holds(C, right_to_use, strongly_true) and holds(A, right_to_allocate, strongly_true)

if receives_no_rights_from(A, C, strongly_true).

 

 

 

 

 

 

 

References Cited

 

 

Benfer, R. A. Jr., Edward E. Brent and Louanna Fubee

1991 Expert Systems. Newbury Park: Sage Publications, Inc.

 

Hamill, James F.

1990 Ethno-Logic. Urbana: University of Illinois Press.

 

Hutchins, Edwin

1980 Culture and Inference. Boston: Harvard University Press.

 

Malinowski, Bronislaw

1935 Soil-Tilling and Agricultural Rites in the Trobriand

Islands. Bloomington: Indiana University Press.

 

Pfaffenberger, Bryan

1988 Microcomputer Applications in Qualitative Research.

Newbury Park, Ca.: Sage Publications, Inc.

 

Weiner, Annette

1976 Women of Value, Men of Renown: New Perspectives in

Trobriand Exchange. Austin: University of Texas Press.