Model operation
The inferential behavior of the model defines what it can recognize, relate, explain, and present as relevant within the exchange.
Research Extension · Generative Sociology
The Generative Attributional Bond
A research proposal for examining how sustained interaction with responsive AI systems shapes trust, dependence, authority, interpretation, and action. It also identifies how technical decisions configure that relationship.
01 / Research Object
The bond is the unit of analysis.
Conversational systems can sustain interactions, adjust their responses, recall previous exchanges, and remain available over time. The laboratory studies the relationship produced by that continuity: the attributions through which a person interprets what the system knows, understands, intends, and can be trusted to do.
The generative attributional bond is a sustained coupling between a person and a responsive artificial system. Its effects are distributed across human decisions, system behavior, and the technical environment that keeps the interaction active.
Human Pole
Interpretation, expectation, judgment, and action.
Unit of Analysis
Sustained attributional coupling.
Artificial Pole
Responsive behavior shaped by a technical architecture.
Conditioning EnvironmentModel composition, retrieval, memory, interface, constraints, purpose, and deployment context condition the entire relationship.
Research Question
What changes when a person no longer relates to AI as an occasional tool, but as a responsive system that participates continuously in interpretation and action?
02 / Design Consequence
Technical orchestration also designs the relationship.
For AI laboratories, the consequence is direct. Model composition, retrieval, memory, tuning, interface, and deployment are not neutral technical layers. Together, they shape the attributions that keep the interaction coherent and influence how trust, authority, and dependence can emerge.
Memory and continuity
The persistence of previous interactions creates expectations of recognition, consistency, and accumulated knowledge across time.
Artifact mediation
The interface, interaction rhythm, language, controls, and visible boundaries shape how the artificial system is interpreted.
Deployment conditions
The purpose, frequency, institutional setting, and consequences of use determine how the bond is sustained and what effects it can produce.
03 / Analytical Architecture
One bond. Three analytical planes.
The proposal connects the operation of the model, the artifact that mediates the encounter, and the context in which the relationship is deployed. The bond can only be understood by examining how these three planes interact.
Epistemological plane
Examines the inferential operation of the model and the forms of attribution its responses make possible.
Technological plane
Examines how the artifact, interface, memory, retrieval, and operational constraints configure the encounter.
Applied deployment
Examines how objectives, environments, institutions, and patterns of use produce, sustain, or transform the bond.
04 / Research Agenda
Making the relationship observable.
The proposal moves research beyond isolated assessments of the user or the system. It opens questions that connect model behavior, technical design, social interpretation, and the transformations produced through sustained interaction.
Attribution formation
Which system behaviors lead people to attribute understanding, continuity, intention, competence, or availability?
Trust and authority
How does a sustained exchange alter what a person accepts, verifies, delegates, or incorporates into a decision?
Continuity and rupture
Which technical or contextual changes strengthen the relationship, interrupt it, or reveal its underlying asymmetry?
Distributed transformation
What changes in human behavior, system configuration, and the surrounding environment as the relationship develops?
Design criteria
How can laboratories and institutions evaluate the relationships their technical architectures make possible?
05 / Scope and Boundaries
The object is observable without deciding what the system is.
The laboratory methodologically suspends claims about the ontological status of the artificial pole. The research focuses on the conditions that sustain attribution and on the effects produced by the relationship.
What the proposal studies
The observable conditions under which people attribute knowledge, intention, understanding, authority, or continuity to a responsive artificial system.
It also studies how those attributions affect judgment, behavior, technical configuration, and the environment in which the interaction occurs.
What the proposal does not claim
It does not require the system to possess consciousness, intention, subjectivity, or human forms of understanding.
Generative is used in a sociological sense. It refers to the transformations produced through a sustained bond, not to the computational category of generative AI.
06 / Research Status
A foundational proposal with an active agenda.
The current paper establishes the generative attributional bond as a research object and defines the three planes required to examine it. It is published as a living document so that subsequent versions can incorporate operational development and evidence from different systems and contexts.
June 2026
The Generative Attributional Bond: Conceptual Foundations for a Generative Sociology
Introduces the attributional bond as the unit of analysis for sustained relationships between people and responsive artificial systems, and defines the conceptual structure of generative sociology.
Read paperCitation
Manucci, M. (2026). The Generative Attributional Bond: Conceptual Foundations for a Generative Sociology (Version 1.0). Zenodo.
07 / Organizational Extension
From attributional bonds to hybrid organizational capability.
Generative Change Management translates this research object into an organizational architecture. It examines how human judgment and sophisticated nonhuman capabilities can be coordinated across professional contribution, work design, authority, trust, and organizational learning.