Information Manipulation Theory 2

Information manipulation theory 2 (or IMT2) is a propositional theory of deceptive discourse production, originally presented at the National Communication Association conference in November 2012. Rooted in extant research and theory on cognitive science, speech production, and artificial intelligence models of problem solving, IMT2 explicates the production architecture underlying deceptive discourse and then presents a series of 15 testable propositions.

IMT2 is an elaboration of arguments presented by Steven McCornack in two previous works: his journal article “Information Manipulation Theory” and a chapter on deceptive message production written for Message Production: Advances in Communication Theory, edited by John O. Greene. In his article on IMT, McCornack suggested that when deceiving others, people play with or “manipulate” information along multiple dimensions in myriad ways that align with the conversational maxims proposed by philosopher Paul Grice. However, because such manipulations are covert rather than overt, listeners are misled.

In his book chapter, McCornack proposed that the production of deceptive discourse can best be characterized as a process of opportunistic problem-solving in which individuals reason through difficult situations, and opt to manipulate information as the most efficient solution for resolving gaps between current problematic states and desired end states. Similar to these earlier offerings, IMT2 conceptually frames deception as involving the covert manipulation of information along multiple dimensions, and as a contextual problem-solving activity driven by the desire for quick, efficient, and viable communicative solutions.

The cognitive discourse-production architecture posited by IMT2 derives from six sources: the theorizing of linguist George Zipf; the blackboard model of problem-solving suggested by H. Penny Nii; principles of design theory forwarded by Vinod Goel and Peter Pirolli; Bernard Baars's global workspace theory; Theo Herrmann's Mannheim model of speech production; and the work of Gary Dell, Franklin Chang, and Zenzi Griffin on lexical encoding patterns.

From these sources, six foundational premises about the nature of the mind, discourse production in general, and deception in particular undergird IMT2: (1) given its frequency, deceptive discourse must present cognitive efficiency advantages over truth-telling; (2) deception derives from opportunistic problem-solving; (3) message design decisions that guide communicative output are driven largely by memory retrieval, not top-down reasoning; (4) deceptive discourse—like all discourse—is not produced as discrete, unitary “messages” in a sequential, linear fashion, but instead results from a series of “microbursts” of mental activity, consisting of cyclical processing, as consciousness interacts with unconscious, modular subsystems including memory, motivations, and speech production; (5) discourse production begins with the construction of a cognitive representation of the situation, which activates relevant information in memory; and (6) the production of spoken discourse occurs incrementally, with the “front end” of utterances being spoken as the “tail ends” are still being designed.

Synthesizing these disparate influences into a coherent whole, IMT2 suggests a multistage architecture of deceptive discourse production. First, a cognitive representation of the “current state” is constructed that activates associated information within memory, in a spreading-activation fashion, feeding a “propositional base” of information potentially relevant for disclosure. Second, this situational representation is compared [Page 510]against a desired end state, based upon conscious and unconscious goals, motives, and intentions. Third, an “action course” through the “problem space” is plotted, based largely on efficiency of solution. Fourth, truthful information within the propositional base that efficiently resolves current-state/end-state discrepancies is fed to speech production in the form of lemmas (nonphonological cognitive representations of words), that are then translated into phonological form and output as spoken discourse. Fifth, current-state/end-state discrepancies that cannot be resolved efficiently through disclosure of truthful information will be resolved by suppressing the information, veiling it through ambiguous word forms, or presenting false information. Sixth, in instances of false-information disclosure, information will be fed to speech production that is most easily accessible, based on memory activation deriving from the cognitive representation of the situation.

...