Alongside the books, this work develops a parallel research program in longitudinal text analysis, narrative structure, and AI-assisted interpretation. The central question is simple but under-addressed:
How do human beings change across time in language, and how can that change be studied without collapsing lives into static labels or one-off snapshots?
This research sits at the intersection of psychology, psycholinguistics, narrative analysis, and large language models. It focuses not on personality typing or automated diagnosis, but on the more difficult problem of trajectory: drift, rupture, reconfiguration, reflective development, and change across extended text corpora.
To examine how narrative, emotional expression, stance, and identity-related signals change across time rather than at a single moment.
To compare lexicon-based, supervised, embedding-based, and LLM-assisted approaches, and to clarify where each behaves as a sensor, an interpreter, or a source of distortion.
To explore how these methods may eventually support broader systems for reflective monitoring, cultural identity research, reader interaction, and structured dialogue analysis.
This work is not driven by the assumption that AI “understands” people in any human sense. It is driven by a narrower and more useful proposition: that structured text, observed over time, contains patterns that can be made more legible through careful methodological design.
Most text analysis still treats language as a snapshot. A text is sampled, scored, classified, and reduced to a single moment of inference. But many of the most important human questions are not about static states. They are about movement.
How does a person's emotional language shift across months or years?
How does reflective capacity deepen or flatten?
How do identities drift across environments?
What changes when language is tracked as a trajectory rather than a set of isolated fragments?
The work presented here approaches language as a longitudinal signal. It asks not only what a text contains, but how its recurring patterns change, reappear, stabilize, or fragment across time and context. This includes work on:
The broader aim is methodological: to help build tools and frameworks capable of studying change in naturalistic human text without pretending that all relevant structure can be captured by single-score outputs.
A framework paper introducing Fourth Culture as a model for understanding culturally hybrid identity in the age of AI-assisted reflection and integration.
Status: Open preprintThis study shows that within-person communication shifts systematically across environments. Using naturalistic text from multiple relational contexts, it argues that drift is contextual rather than evidence for trait inference, and that the resulting axes should be treated as analytical scaffolds rather than personality claims.
Status: Open preprintUsing a single-subject longitudinal corpus, this paper argues that existing tools often capture surface variation while failing to infer stable longitudinal psychological trajectory. It highlights message-length confounding and the limitations of treating psychological change as a straightforward extraction problem.
Status: Open preprintA validation framework for emotion extraction that distinguishes textual expression from reported felt state, showing that disagreement between methods can be informative rather than merely erroneous.
Status: Open preprintThis paper uses sustained human-LLM dialogue to characterize macro-scale dynamics across time. It focuses on embedding generation, clustering, entropy, divergence, and coupling metrics to capture change in long-form dialogue systems.
Status: Open preprintA conceptual and protocol-oriented paper exploring whether a structured AI-assisted reflective framework can generate longitudinal material suitable for identity research and integration in culturally hybrid lives.
Status: Open preprintThe Global Narrative Atlas (GNA) is the most ambitious element of the research program. It is not simply a paper, and not yet a full platform. It is best understood as an emerging research infrastructure concept: a framework for studying how identity, emotion, and reflective structure evolve across time in naturalistic text.
At its core, GNA is based on a simple observation: current research captures many snapshots of human psychology, but far fewer trajectories. We know a great deal about isolated states and cross-sectional differences, and much less about how identities actually change as people move through migration, instability, technological mediation, and long-form reflection.
GNA is intended as one response to that gap.
The research presented here takes a deliberately cautious stance toward AI methods.
Large language models are powerful, but they are not neutral. They can behave as flexible interpreters, useful comparators, and structured summarizers, but they can also hallucinate, over-smooth, and conceal uncertainty behind fluent language. Lexicons, classifiers, embedding systems, and prompted LLMs should therefore not be treated as interchangeable solutions. They occupy different methodological roles.
One of the recurring themes of this work is that disagreement between methods is often informative. It can reveal construct boundaries, interpretive limits, and differences between sensing and explanation. The aim is not to find one magical model, but to build a saner methodological language for working with longitudinal human text.
This research does not stand apart from the books and applied systems. It exists in dialogue with them.
The books provide conceptual terrain: identity, belonging, emotional architecture, reflective life.
The research attempts to make aspects of that terrain analytically tractable.
The applied systems explore what happens when these ideas are moved into reader interaction, structured dialogue, and conversational assessment.
In this sense, the research program is not a detached technical layer. It is one part of a larger inquiry into how narrative, identity, and reflection can be understood across time in the age of AI.
For readers coming from the books, the research pages offer the methodological side of the same broader inquiry. For researchers, they offer a view of how narrative systems, identity theory, and longitudinal methods might be brought into closer contact.
