pith. sign in

arxiv: 1907.09293 · v1 · pith:WUESHOOSnew · submitted 2019-07-19 · 💻 cs.AI · cs.CL· cs.HC· cs.MA· cs.MM

DREAMT -- Embodied Motivational Conversational Storytelling

David M W Powers This is my paper

Pith reviewed 2026-05-24 19:43 UTC · model grok-4.3

classification 💻 cs.AI cs.CLcs.HCcs.MAcs.MM
keywords storytellingembodied conversational agentsDREAMT modelcharacter-driven systemsmotivational modellingeducational interventionsaccessible interfacesconversational storytelling
0
0 comments X

The pith

The DREAMT model identifies six layers required for character-driven storytelling in embodied conversational agents.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that storytelling is fundamental to language and thus emerges as an essential component of intelligent systems, including embodied conversational agents used in education, health interventions, and accessible interfaces for people with disabilities. It characterizes storytelling as an inventive fleshing out of detail from a particular personal perspective. The central proposal is the DREAMT model, a mnemonic framework that organizes the necessary layers in such systems. This model draws from the authors' prior implemented research but is presented as a way to focus attention on integrating the components more fully. The layers cover description and dialogue, realization and role, explanation and entertainment, actualization, motivation and modelling, and topicalization and transformation.

Core claim

Storytelling is fundamental to language, including culture, conversation and communication in their broadest senses. It thus emerges as an essential component of intelligent systems, including systems where natural language is not a primary focus or where we do not usually think of a story being involved. We further present a characterization of storytelling as an inventive fleshing out of detail according to a particular personal perspective, and propose the DREAMT model to focus attention on the different layers that need to be present in a character-driven storytelling system, formalized mnemonically as Description/Dialogue/Definition/Denotation, Realization/Representation/Role, and so on

What carries the argument

The DREAMT mnemonic, which breaks storytelling into six paired layer categories that together structure character-driven systems: description/dialogue/definition/denotation, realization/representation/role, explanation/education/entertainment, actualization/activation, motivation/modelling, and topicalization/transformation.

If this is right

  • Storytelling components must be present in embodied agents for effective educational and health interventions.
  • General-purpose interfaces for users with disabilities require character-driven storytelling built on the DREAMT layers.
  • The model applies to systems where natural language is not the main focus.
  • Most aspects of the DREAMT layers have already appeared in primitive form in prior implemented research systems.
  • Formalizing the layers enables better future integration of description, motivation, and transformation elements.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The model could be tested by building two versions of an agent, one with explicit DREAMT layers and one without, then measuring user engagement in a health intervention task.
  • Motivation and modelling layers suggest a direct link to goal-directed behavior that might map onto reinforcement learning objectives.
  • Topicalization and transformation could be examined as mechanisms for how agents adapt stories over multiple interactions.
  • The framework might extend beyond embodied agents to any conversational system where maintaining a consistent character perspective improves coherence.

Load-bearing premise

Storytelling is fundamental to intelligent systems even when natural language is not a primary focus.

What would settle it

An embodied conversational agent that successfully supports education or health interventions without incorporating the DREAMT layers or any storytelling components.

read the original abstract

Storytelling is fundamental to language, including culture, conversation and communication in their broadest senses. It thus emerges as an essential component of intelligent systems, including systems where natural language is not a primary focus or where we do not usually think of a story being involved. In this paper we explore the emergence of storytelling as a requirement in embodied conversational agents, including its role in educational and health interventions, as well as in a general-purpose computer interface for people with disabilities or other constraints that prevent the use of traditional keyboard and speech interfaces. We further present a characterization of storytelling as an inventive fleshing out of detail according to a particular personal perspective, and propose the DREAMT model to focus attention on the different layers that need to be present in a character-driven storytelling system. Most if not all aspects of the DREAMT model have arisen from or been explored in some aspect of our implemented research systems, but currently only at a primitive and relatively unintegrated level. However, this experience leads us to formalize and elaborate the DREAMT model mnemonically as follows: - Description/Dialogue/Definition/Denotation - Realization/Representation/Role - Explanation/Education/Entertainment - Actualization/Activation - Motivation/Modelling - Topicalization/Transformation

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 1 minor

Summary. The paper argues that storytelling is fundamental to language and communication and thus an essential component of intelligent systems, including non-language-focused embodied agents. It proposes the DREAMT mnemonic (Description/Dialogue/Definition/Denotation; Realization/Representation/Role; Explanation/Education/Entertainment; Actualization/Activation; Motivation/Modelling; Topicalization/Transformation) to organize the layers needed in character-driven storytelling systems, claiming that most aspects have arisen from the authors' prior implemented but primitive and unintegrated research systems.

Significance. If the DREAMT framework can be operationalized, it may help organize design considerations for storytelling in embodied conversational agents used in education, health, and accessibility applications. The contribution is primarily conceptual and mnemonic rather than providing new empirical results, formal derivations, or validated implementations.

minor comments (1)
  1. [Abstract] Abstract: the claim that 'Most if not all aspects of the DREAMT model have arisen from or been explored in some aspect of our implemented research systems' would be strengthened by adding specific citations or brief descriptions of the relevant prior systems and their mappings to DREAMT layers.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive review and recommendation of minor revision. The report accurately characterizes the paper as a conceptual contribution centered on the DREAMT mnemonic framework derived from prior work. No specific major comments were listed under the MAJOR COMMENTS section.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper presents DREAMT purely as a mnemonic proposal derived from the author's prior (explicitly primitive) implemented systems. No equations, fitted parameters, predictions, uniqueness theorems, or derivations are claimed. The model is framed as an organizational tool based on experience rather than a forced result from self-citation or self-definition. The mention of prior work is motivational and does not reduce the central claim to its inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The paper is a high-level conceptual proposal; the central claim rests on domain assumptions about the necessity of storytelling rather than on new parameters or entities with independent evidence.

axioms (1)
  • domain assumption Storytelling is fundamental to language, including culture, conversation and communication in their broadest senses.
    Invoked in the first sentence of the abstract as the basis for requiring storytelling in intelligent systems.

pith-pipeline@v0.9.0 · 5756 in / 1089 out tokens · 19870 ms · 2026-05-24T19:43:23.149490+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

17 extracted references · 17 canonical work pages

  1. [1]

    and PARRY (Colby et al. 1972). The ELIZA ‘doctor’ and the PARRY ‘patient’ themselves reflect two sides of the coin we examine in this paper, which focuses on the roles played by agents. The AIML-based ALICE system (Wallace, 1995/2008) and successful conversational/Q&A bots in general (Shah et al.,

    work page 1972

  2. [2]

    still rely largely on templates, with AIML providing an open Artificial Intelligence Markup Language that facilitates writing Talking Head ECAs and toolboxes (Luerssen et al., 2010). We follow in this tradition while downplaying the syntax in favour of semantics – it is very easy to program a sophisticated grammar and use it for generation, and a generati...

    work page 2010

  3. [3]

    or even computer games (Riedl, et. al, 2011), storytelling has a role that is perhaps not so obvious in everyday conversation, and in interactions with students or patients, but we argue that thinking in terms of storytelling offers the opportunity to improve the role of AI in such ECA interventions – both in terms of conveying content in a way that is in...

    work page 2011

  4. [4]

    This involves understanding the user’s character and situation and building an appropriate story

    is the development of Unconscious AudioVisual and Brain Computer Interfaces (UCI/ABCI), with a particular focus on educational and health interventions, including an audiovisual brain controlled wheelchair and a multimodal authoring system for people with disabilities – in both cases these are high-level model-driven systems where the computer is aiming t...

    work page 1998

  5. [5]

    as well as subsuming both taxonomic and meronymic relationships in a general logical spatiotemporal representation. Pike’s broad use of the Tagmemic formalism also suggests how we can extend this common representation principle to cover the modeling and representation of character and reader through the use of dynamic character attributes. We follow Hayes...

    work page 1979

  6. [6]

    The issue of role is critical in storytelling as well as in traditional linguistics (providing a semantic counterpart to the grammatical slot)

    Multirole ECAs: teacher/student/peer/... The issue of role is critical in storytelling as well as in traditional linguistics (providing a semantic counterpart to the grammatical slot). For a traditional ECA, the ECA plays a single role, but in our educational and health training applications, we have found it useful to be able to play multiple roles (usua...

    work page 2000

  7. [7]

    I’ve got an idea…

    and generalizing to stories, discourse and human behaviour in general (1964,1967,1973,1982,1986). Tagmemic grammar characterizes the traditional non-terminals and rewrite rules of phrase structure grammar (PSG) in terms of slot and (filler) class, augmenting them with attributes denoted role and cohesion that fit well within an attribute grammar or unific...

    work page 1964

  8. [8]

    pages 326-337 https://doi.org/10.1007/3-540-54563-8_96

    Springer, Berlin, Heidelberg. pages 326-337 https://doi.org/10.1007/3-540-54563-8_96. Sean P Fitzgibbon, Dylan DeLosAngeles, Trent W Lewis, David MW Powers, Emma M Whitham, John O Willoughby, Kenneth J Pope

    work page doi:10.1007/3-540-54563-8_96

  9. [9]

    Artificial Intelligence 40(1-3):63-118

    Explanation-Based Learning: A Problem Solving Perspective. Artificial Intelligence 40(1-3):63-118. https://doi.org/10.1016/0004-3702(89)90047-7. Kenneth L Pike

    work page doi:10.1016/0004-3702(89)90047-7

  10. [10]

    University of Michigan Press, Ann Arbor

    Phonemics: a Technique for Reducing Languages to Writing. University of Michigan Press, Ann Arbor. Kenneth L Pike. 1964 Discourse analysis and tagmeme matrices. In Elmer Wolfenden (ed.), Papers on Philippines Linguistics by members of the Summer Institute of Linguistics. Oceanic Linguistics 3(1):5-26. Kenneth L Pike

    work page 1964

  11. [11]

    Mouton, The Hague

    Language in Relation to a Unified Theory of the Structure of Human Behavior. Mouton, The Hague. Kenneth L Pike 1973 Science fiction as a test of axioms concerning human behavior. Parma Eldalamberon 1(3):6-7. 11 Kenneth L. Pike and Evelyn G. Pike

    work page 1973

  12. [12]

    Summer Institute of Linguistics and University of Texas, Arlington

    Grammatical Analysis. Summer Institute of Linguistics and University of Texas, Arlington. Kenneth L Pike 1986 A further note on experimental clauses in discourse. In Benjamin F. Elson (ed.), Language in global perspective: Papers in honor of the 50th anniversary of the Summer Institute of Linguistics, 1935-1985, 135-138. Dallas: Summer Institute of Lingui...

    work page 1986

  13. [13]

    Cambridge University Press, Cambridge, UK

    Building Natural Language Generation Systems. Cambridge University Press, Cambridge, UK. Mark Riedl, David Thue, & Vadim Bulitko, (2011). Game AI as Storytelling. In P. A. González-Calero & M. A. Gómez-Martín (Eds.), Artificial Intelligence for Computer Games (pp. 125-150). Springer: doi:10.1007/978-1-4419-8188-2_6 Marie-Laure Ryan, 2003, On Defining Narr...

    work page doi:10.1007/978-1-4419-8188-2_6 2011

  14. [14]

    1 7 ( 1 ): 49-55

    Concept Development for Expert System Knowledge Bases, The Australian Computer Journal, Vo l . 1 7 ( 1 ): 49-55. Huma Shah. 2011a. Deception-detection and machine intelligence in Practical Turing tests. PhD Thesis. University of Reading. Huma Shah. 2011b. Turing’s misunderstood imitation game and IBM’s Watson success. Invited Keynote: 2nd Towards a Compre...

    work page 2011

  15. [15]

    Computers in Human Behavior, vol

    Can Machines Talk? Comparison of Eliza with Modern Dialogue Systems. Computers in Human Behavior, vol. 58 pages 278–295. http://dx.doi.org/10.1016/j.chb.2016.01.004 Luc Steels,

    work page doi:10.1016/j.chb.2016.01.004 2016

  16. [16]

    https://doi.org/10.1186/s40469-016-0008-2

    Mimicry and expressiveness of an ECA in human-agent interaction: familiarity breeds content! Computational Cognitive Science 2 (1):1. https://doi.org/10.1186/s40469-016-0008-2. Andrea Tartaro and Justine Cassell

    work page doi:10.1186/s40469-016-0008-2

  17. [17]

    Proceedings of International Conference of the Learning Sciences

    Playing with Virtual Peers: Bootstrapping Contingent Discourse in Children with Autism. Proceedings of International Conference of the Learning Sciences. Alan M. Turing (1950) Computing Machinery and Intelligence. Mind 49:433-460. Richard Wallace

    work page 1950