Kaufer-Carley Theory of Engineered Rearch

Betsy V. Martens bvmarten at MAILBOX.SYR.EDU
Mon Aug 14 09:15:11 EDT 2000

> Date:    Sat, 12 Aug 2000 12:07:41 +0200
> From:    "Michel J. Menou" <Michel.Menou at WANADOO.FR>
> Subject: Re: Controversial Questions
> At 15:05 11/08/00 -0400, Betsy Van der Veer Martens wrote:
> >snip
> >Just as additional conversation fodder: I'd be curious myself to know
> >what people on the list think of the theory of "engineered reach" as
> >applied to citations by David Kaufer and Kathleen Carley in their 1993
> >book, "Communication at a Distance"? Is anyone attempting to replicate
> >their simulated findings with actual citation data? Why or why not?
> As the idiot on duty in this list, I'd also be interested to see a short
> introduction to this topic and/or a reference.
> Thanks in advance
> Michel Menou
Dear Michel, Gene, and interested others,

Alas, I certainly don't travel in those rarefied Carnegie-Mellon circles
with Drs. Kaufer and Carley, and only came across their work in the course
of reviewing the literature for my own dissertation in progress. But
following is a brief description of the Carley-Kaufer "theory of
engineered reach", omitting the mathematical elements of their model,
which the book addresses in detail. Please forgive the embedded references
and the doctoral student tone: this is taken directly from my lit review.
Perhaps others on the list with more knowledge of this work could comment?


Betsy Martens
Syracuse University
School of Information Studies
bvmarten at mailbox.syr.edu

Kaufer, David S. and Carley, Kathleen M. (1993) COMMUNICATION AT A
Hillsdale NJ: Lawrence Erlbaum Associates.

In their 1993 monograph, "Communication at a Distance", David Kaufer and
Kathleen Carley propose a systems approach to the communicative
transaction, integrating agents, content, and context as components of a
sociocultural ecology which evolves over time (Kaufer and Carley, 1993,
pages 87-88). This unified theory of the communicative interaction cycle,
which they term "constructuralism" (Kaufer and Carley, 1993, page 14), is
suggested as a more satisfactory meso-theoretical approach than either of
the two pre-eminent communication research traditions: the empirical
(e.g., mass communications theory) and humanistic (e.g., reader-response
theory). Both of these separate traditions tend to make either the
specific communicator, the audience, the communication, the communication
technology, or the communicative effect the object of inquiry. Both
traditions, therefore, tend to "black-box" the communicative transaction
itself (Kaufer and Carley, 1993, page 88). Further, while the empirical
communication research tradition functions ahistorically, in accordance
with the social scientific research tradition from which it stems, the
humanistic research tradition is historically specific in its concern with
individual authors, in accordance with the classic literary research
tradition (Kaufer and Carley, 1993, page 2). Consequently, both traditions
neglect the continuities of social and cultural organization arising from
the circulation of information through concurrent transactions unfolding
over time (Kaufer and Carley, 1993, page 3). In contrast, the
constructuralist approach, which "... makes individual cognition
(specifically, individual information; the ability to learn and to choose
interaction partners with whom to learn) the central showpiece [in which]
the slightest change in what an individual knows can, in principle, affect
the entire sociocultural landscape", also allows the aggregation of the
communicative transactions which dynamically structure a given social
world over time (Kaufer and Carley, 1993, page 90).

Kaufer and Carley point out the critical centrality of print in modern
sociocultural organization and change, appreciating in particular the
influence of such predecessors as Eisenstein (1979), Goody (1986), Innis
(1950, 1959), McLuhan (1962, 1964), Ong (1982), and Yates (1966) in
exploring the impact of old and new media on basic cognitive and cultural
organization for understanding information, from units organized by the
oral/aural senses to units organized by the visual senses (Kaufer and
Carley, 1993; page 9). Similarly, they acknowledge the importance of such
social theorists as Archer (1988), Callon and Latour (1981), Collins
(1975), Giddens (1984), and Turner (1988), all of whose work explore in
various ways the "meso-structure" between the micro-interactions of
individuals and the macro-structures of society (Kaufer and Carley, 1993,
page 206). However, constructuralism itself can be profitably viewed as a
structural theory of information diffusion, with ties to classic work in
diffusion theory (Katz et al., 1963; Rogers, 1962), information use theory
(Feldman and March, 1981; Festinger, 1956), social network theory (Burt,
1983; Granovetter; 1973; Pool and Kochen, 1978; White and Breiger, 1975),
and scientometric theory (Holton, 1978; Price, 1970).

Constructuralism provides a dynamic model of the communicative transaction
cycle as it impacts and is impacted by the particulars of any
sociocultural topology. It specifically foregrounds "adaptation" by social
groups in consonance with its emphasis on change over time, with three
specific "measures of adaptation" at the sociocultural level: diffusion,
stability, and consensus (Kaufer and Carley, 1993, page 148). Diffusion is
a longitudinal measure of the fraction of the population that has received
some new information or innovation. Stability is the degree to which the
population shares available information. Consensus is the degree to which
the population shares the same belief about some focal idea or decision

Constructuralism, thus, is a generic process model that Kaufer and Carley
employ in a series of historically-informed, computer-assisted thought
experiments that model the movement of texts as artificial
agents in simulated societies of oral and print-oriented communicators,
professional communicators, and academic communicators. As the intent of
these simulations is theory development, framed by high-level historical
accounts and the operationalization of these accounts into a set of models
that furnish insight into how the accounts need further elaboration and
revision, constructuralism is very much a work in progress (Kaufer and
Carley, 1993, page 8).

In a critique of self-organizing systems research in the social sciences,
including both constructuralism and autopoiesis, Noshir Contractor points
out that such models run the risk of being more metaphorically appealing
than theoretically or empirically revealing (Contractor, 1998). The
considerable difficulties in modeling communicative complications are
exemplified in Niklas Luhmann's attempt to use autopoiesis to model
communication systems such as law (Luhmann, 1982, 1995). Autopoiesis has
made little progress to date in modeling a real-world legal knowledge
system or in fitting actual or gedanken data to its parameters (Baxter,
1998; DeFlem, 1998).

One of the more fascinating features of the constructuralist model is the
specification of the text as an artificial agent encapsulating part of its
author's mental model during the time of composition, and subject to
later, local interpretation during communicative transactions (Kaufer and
Carley, 1993, page 229). In the academic simulation, this specification is
termed the "theory of engineered reach" and makes a distinction between
context-related (prestige of author and/or journal of publication) and
content-related resources available to authors (Kaufer and Carley, 1993,
pages 380-1). Authors' content-related resources for creating their
textual agents consist of "citation" (the use of authors' names to relate
the textual content to the ideas of other authors in the field) and
"elaboration" (the use of sentences to expand more fully than citations
upon the ideas of prior texts). Kaufer and Carley quantify both citation
and elaboration in a test of their "engineered reach" model (Kaufer and
Carley, 1993, pages 385-391). The neglect of citation and elaboration
quality in their model presumably contributes to the inconclusiveness of
their results.

What is not at present factored into constructuralism's textual agent is
any potential power or truth-value of those new ideas embodied in the text
itself. While avoiding the danger of a simplistic evolutionary
epistemology such as Dawkin's concept of "memes" (self-validating ideas as
controlling parasites on the brains of their human hosts) which has been
recently been both popularized (Balkin, 1998; Blackmore, 1998; Lynch,
1996) and debunked (Gatherer, 1998; Polichak 1998), the textual agent
comes precariously close to foundering on a purely social construction of
truth as an emergent property of subsequent local interpretations. It is
unlikely that the current model of constructuralism would be able to
explain much about the actual import and impact of the ideas of, for
example, Darwin (Dennett, 1995) or even Derrida (Lamont, 1987) without
attending to larger issues of private or public standards for credibility.

Private standards for credibility would presumably involve individually
distinguishing the ontology of agents (e.g., as facts, hypotheses,
assumptions, speculations, etc.). Public standards for credibility might
be Bayesian, possibly using Goldman's (1999) "veritistic social
epistemology", which allows for varying degrees of belief to be based on
the range of proven or possible "veritistic values" of a particular item
of information. "Consensus consequentialism" (that is, mistaking the
consensus of any group on a particular matter as equivalent to the "truth"
of that matter) inherent in the constructuralist model at present could
perhaps be averted by allowing credibility to be set at different levels
for different agents and audiences. This Bayesian approach, however, is
itself imperfect and attests to the difficulty of modeling credibility
claims in even a simulation of knowledge transfer and transformation.

A further weakness of the constructuralist model is that it tends to
present the communication landscape as a level plain, without the
complications introduced by the dynamics of various knowledge systems and
their interrelationships. Although this is an inherent feature of models,
advances in constructuralist research beyond what is already available
through standard diffusion research or network analytic techniques will
obviously require the introduction of appropriate cognitive and
communication complications.

Although constructuralism's use of "citation" as textual agent may be
particularly problematic, it does possess the inestimable advantage of
providing a useful framework for empirical research. Clearly,
constructuralism's strengths as an integrated approach to the study of the
development of reliable knowledge, as compared with its predecessors, are
well worth further discussion and elaboration. In order to move beyond
thought experiments, though, the model needs to be supplied with empirical
data from a variety of communicative perspectives.

More information about the SIGMETRICS mailing list