Miguel Ángel Ballesteros bio photo

Miguel Ángel Ballesteros

Maker, using software to bring great ideas to life. Manager, empowering and developing people to achieve meaningful goals. Father, devoted to family. Lifelong learner, with a passion for generative AI.

Email LinkedIn Github
RSS Feed

When Text Sounds Like Me, I Understand It Better

Published on and last modified on .

Available in Español .

It happened to me with the epigenetics article: the papers were impeccable, yes, but the rewritten version in “my voice” (generated by AI from my texts) was infinitely clearer to me. My turns of phrase, my order, my connectors. Suddenly, everything clicked.

Simple hypothesis: if you adapt the style of the material to the style of the reader, comprehension increases. Not because “I generated it” (there is no generation effect here), but because the text fits my schemas, sounds like my head, and aligns with how I process language. That has a basis.

Why it (yes) works

  • Schema Congruence. We understand better when the new fits into previous structures (schema theory: Rumelhart). A text that respects my typical way of organizing ideas reduces the integration effort. (academypublication.com)
  • Processing Fluency. What is processed more fluently is judged to be clearer and “truer”. If the text uses my patterns (sentence length, connectors, rhythm), fluency increases and the feeling of comprehension as well. Note: feeling ≠ learning (I’ll come back to this). (SAGE Journals)
  • Self-Reference. When the content “speaks to me” in my own register, it touches the self and is encoded better (self-reference effect). (PubMed)
  • Linguistic Alignment. In real dialogue, we tend to synchronize representations at multiple levels; this alignment facilitates understanding. A text “in my voice” simulates part of that effect. (Cambridge University Press & Assessment)
  • Discourse Personalization. Research in multimedia shows that a conversational/personal style improves learning compared to an impersonal/formal one. Adapting voice and tone is not cosmetic: it is instructional. (Cambridge University Press & Assessment)

In short: I’m not talking about “generating to remember more”, I’m talking about congruence: when the form of the text resonates with my schemas and my style, I understand better now… and I can remember better later if I combine it with the right practice.

“Okay, what if I apply it with my high school nephew?”

A quick, concrete, and measurable proposal:

  1. Style Capture. Gather 1,000–2,000 of his words (letters, stories, papers). Extract signals: average sentence length, frequent connectors (“so”, “in short”), level of technicality, typical examples.

  2. AI Transformation. Feed the syllabus (Economics / G&H) and ask the AI for two versions:

    • Official (style untouched).
    • “Nephew” version (same information, his voice).

  3. A/B Test with Delay.

    • Study one unit with each version.
    • 48–72 hours later, free and applied recall test (2–3 problems). The metric is better retention and transfer in the “Nephew” version. (PubMed)
  4. Iterate. Adjust connectors, examples, and metaphors that worked best; repeat in another unit.

Risk and Antidote: The “Illusion of Fluency”

A known danger: when something flows too much, I believe I know it better than I actually do (illusions of competence). Solution: combine the personalized text with retrieval practice (spaced mini-tests, explaining aloud, etc.). Testing > rereading. (PubMed)

Closing

My discovery is not “write it yourself and you will learn more”, but have it explained to you in your own voice. That layer of transformation —including AI— reduces friction, aligns schemas, and brings the concept closer. Then, you finish it off with spaced retrieval to consolidate. Immediate clarity + long-term memory. (academypublication.com)

When Text Sounds Like Me, I Understand It Better was published on and last modified on .