К оглавлению
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 
136 137 138 139 140 141 142 143 144 145 146 

The problem space of social causal mapping centers on the use of individual causal maps

to represent and understand social causal cognition. Representing this problem space

involves a variety of abstractions, ranging from lower-level views of the constituent

components of cognitive maps (e.g., nodes and arcs) to higher-level aggregations of

individual maps (e.g., a chain of node-arc-node segments describing a chain of causality

drawn from the assertions of multiple individuals). These abstract structures and their

behaviors constitute examples of object structures of a logical model.

Object structures are often based on a compositional, or “part of” architecture, such as

that found in an automobile composed of an electrical system, a power system, an air

conditioning system, and so forth. In turn, the electrical system is composed of a fuse

system, a wiring system, and a battery system. The power system is composed of an

engine, a fuel system, a lubrication system, and a cooling system. This compositional

architecture also applies to causal and other cognitive maps. Like an automobile, a causal

map can be viewed as a single object and as an integrated system of component parts.

Just as automobiles can be envisioned as a composition of an engine, brakes, wheels,

headlights, etc., so too can causal maps be envisioned as a composition of nodes and

arcs. Social causal maps can be envisioned as a composition of individual causal maps

(e.g., Figure 2), or as a composition of cognitive maps (Figure 1).

This compositional view is called aggregation, “an abstraction which allows a relationship

between named objects to be thought of as a (higher-level) named object” (Smith

& Smith, 1977a). Compositional views of object structure are typically designed according

to varying levels of granularity within the problem space of social causal mapping.

At the fundamental level, “fine-grained” objects representing nodes and arcs are

aggregated into higher-level “coarse-grained” objects (e.g., causal, categorical, or

associative maps) that represent an integration of the individual’s assertions about his

or her thoughts and beliefs. In turn, these individual cognitive maps can be organized

into still higher-level aggregations, such as the linking of segments of cognitive maps

from multiple individuals, where “relevant” may encompass entire cognitive maps, or if

need be, only those portions that meet a specific characteristic (e.g., those that form a

path between two nodes of interest). The object’s granularity (i.e., its breadth and scope)

is determined largely by the issue at hand and the way the users have framed that issue

within the problem space.

Object structures can be described in terms of their data structure and methods. These

two descriptions will now be explored in greater detail, particularly in regard to the way

in which they support information technologies for social causal mapping.

Data Structures for Social Causal Mapping

Data structures provide the means for storing information about critical elements of a

problem space. In causal maps, nodes represent such things as physical items or mental

concepts, and are typically described with the name of that item or concept. Arcs

represent assertions about causal relationships between nodes. They have a distinct

“head” and “tail,” wherein the “head” of the arc points to the node that is influenced by

the node at the “tail” of the arc. Arcs are typically valued as “+” or “-” to denote whether

the arc represents a directly or inversely proportional relationship, respectively. These

various characteristics (e.g., node names, arc directions, arc “+” or “-” values) are

outlined in data structures, in that the data structures provide blueprints for the

construction of data storage.

Data structures should also contain information to help users of social causal mapping

systems obtain deeper understandings of causal assertions. This information could

include the contributors’ names, contact information (e.g., e-mail address and phone

number) and their departments. It may even include references to simulations or other

decision models that contributors used to develop or justify their beliefs, or to diverse

media files such as text, graphs, pictures, audio, and/or video (Boland et al., 1994). Such

information, when contained in or referenced by the object data structures within a social

causal mapping system, can facilitate organizational learning and knowledge management

when they are designed to link concise abstract knowledge elements embodied in

causal maps to the rich “fluid mix of framed experience, values, contextual information,

and expert insight” of their organizational participants. The addition of data structures

storing contributors’ names or identification numbers can implement this linkage, thus

allowing collections of causal maps to “map” the organizational knowledge “terrain.” The

use of such labels requires that members within that terrain must be allowed to contribute

knowledge freely and with sufficient confidence so that they will willingly attach their

names or employee identification numbers to their contributions. Then and only then can

the chain of cognitive patterns, contributing individuals, and their rich tacit knowledge

be reliably forged into a useful knowledge management tool.

For social causal maps to be an effective organizational learning tool, the information

about those maps and their contributors must be available for long-term use. In objectoriented

modeling, this concern is called object persistence (i.e., how the objects are

stored over time). Most cognitive mapping tools available today typically store this

information in modularized but non-integratable files (usually graphic). This approach

creates difficulties in analyzing large numbers of social causal maps and the social

cognitions they represent. The proposed social causal mapping system differs from other

mapping approaches by stipulating an inclusive, comprehensive system of storage

whose components can be integrated regardless of the time, place, or individuals from

whom they were elicited. This goal of comprehensive storage can be achieved with a

design that enforces the way new cognitive maps are added to the system. Objects can

be stored in one or more long-term structures and media as long as the data and

information within those structures and media are integratable with the rest of the

proposed system.1

Methods for Cognitive Mapping Systems

Causal maps, and the other cognitive maps that can augment their collectivization,

represent important organizational knowledge. This knowledge can be quite useful in

addressing a variety of organizational problems, including those associated with the

analysis, design, implementation, and maintenance of information systems (e.g., improving

customer service, reducing costs, or streamlining business processes). The causal

assertions of organizational members thus form an important but complex part of

organizational memory.

The potential size and complexity of creating, manipulating, storing, and retrieving large

numbers of cognitive maps demands the use of a computer-based information system.

Software components within such a computer-based information system can be employed

by the user to manipulate object structures in a variety of social causal mapping

activities. These components, called methods, can be used to construct and destroy

nodes and arcs, update and maintain data within data structures, traverse chains of nodearc-

nodes, calculate the total effects of node-arc-node chains (e.g., xyz = x z), and retain

cognitive maps in long-term storage. A more generalized collection of functions for

organizational memory information systems (OMIS) has been identified by Stein and

Zwass (1995). Their collection contains five sets of functions that should be applicable

to social causal mapping systems because such systems are a form of organizational

memory. These five sets will now be described and adapted to the design of methods

providing similar functionalities for social causal mapping.

Mnemonic Functions

Mnemonic functions address the “acquisition, retention, maintenance, search, and

retrieval of information” and knowledge (Stein & Zwass, 1995). These functions are

typical of most kinds of information systems. However, retention warrants some discussion

here because of problems that have resulted from the way in which causal maps have

been retained in the past. Retention refers to the way in which nodes, arcs, and related

data should be placed in long-term storage. In the past, individual cognitive maps have

been stored primarily through segregated graphic files. Such files have been difficult to

merge, and often require extensive use of “copy and paste” procedures. Retention design

also affects the design of maintenance functions: in a multi-user system, maintenance

functions must be secure so that contributors are prevented from modifying the maps

of others. A maintenance function would have to be designed so that contributors could

update and modify their cognitive maps while mitigating data inconsistencies (e.g., if a

user changes the name of a node, that change would be reflected in all cognitive maps

using that node). Search and retrieval should be non-problematic.

Integrative Functions

Integrative functions are designed “to ensure that the internal knowledge of the

organization regarding technical issues, past decisions, projects, designs, and so on is

made explicit and available for future use, complete with contexts, rationales, and

outcomes” (Stein & Zwass, 1995). Internal knowledge within organizational memory can

be made available by integrative functions that build chains of contiguous node-arcnode

segments across the causal maps of individuals (e.g., the path g→m→n→e in Figure

2). When those paths contain information that can identify the individuals who have

contributed their causal maps to the database, the user is given the option of locating

those individuals and engaging them in rich social dialogue. Such individuals are more

complete sources of knowledge than the terse nodes contained in their causal maps, and

may offer more insight and flexibility than the textual, graphic, or other repositories in the

organization’s knowledge base.