Decomposing creation through the lens of letter-strings.

How a micro-world invented 30 years ago provides helpful perspectives over today’s copyright challenges in the light of AI.

Riccardo Zanardelli


This brief piece takes inspiration from the approach described by Hofstadter and Mitchell in the making of Copycat, as extensively illustrated by Melanie Mitchell herself in the book “Analogy-Making and Perception” (1993). In the book, Mitchell perfectly arguments the advantages of the micro-world made of letter-strings in the study of high-level perception, conceptual slippage and analogy making. Similar advantages could be beneficial in the analysis of some intellectual property questions in the times of AI -this is my perception.

In her own words:

“The point of working in a microdomain in cognitive science is to isolate a phenomenon (such as analogy making), strip it down to its bare bones, and get rid of its extraneous real-world trappings while returning its essence so that it can be investigated more clearly”

Our stylised little world.

In our micro-world, entities exist only in the form of letter-strings (i.e. “abc”).

In our micro-world, we create new letter-strings by transforming and/or combining existing letters-strings through rules, which we have learned by looking at examples.

In our micro-world, rules are ideal functions mapping one letter space to another, they are learned by analogy and they can be combined in sequences of the desired complexity.

In our micro-world, at some point, one might be willing to determine if the result of our creation, or eventually the chain or “creation rules” themselves, is original and new such that its copyright can be claimed and managed.

Assuming that we have already shared and agreed upon what is a letter, and what is a letter-string, this is enough to describe our world, but let’s make an example for more clarity, the same example opening Mitchell’s book:

  • ( abc ) is a letter-string;



Riccardo Zanardelli

Digital Platforms @ Beretta | PhD student in Statistics & Data Science @ AEM, UNIBS | Engineer | Only personal opinions here | Code is Law (cit.)