The work presented in this article, while constituting an autonomous study, originated from a need identified during the design of a broader experiment aimed at integrating computational thinking in early education. The research aligns with constructivist pedagogy, emphasizing active learning through tangible and narrative-driven experiences. The study involved primary school students and focused on both theoretical prerequisites (such as symbolic representation and rule-based decision making) and technological foundations (basic I/O devices and software interaction).

Using a mixed-methods approach, the experience demonstrated that key computer science concepts -including syntax, state transitions, and abstraction- can be effectively introduced in early primary education through activities based on storytelling, hands-on manipulation, and visual representation. These methods are designed to adapt to developmental stages where concrete-operational thinking dominates, bridging the gap between play and formal logic.

The "language of crystals", a metaphorical gateway for abstract reasoning patterns which students must learn to become "photonic wizards", was embedded in the engaging narrative framework of the "labyrinth of lights" - a multi-sensory artifact where students manipulated reflective mirrors ("crystals") placed along the diagonals of a 3×3 physical grid (the "labyrinth") to intentionally shape light paths using a laser pointer. To convey the rules of such language, reflections were stimulated through analogies with natural language, evoking symbols, words, and sentences with the goal of helping students realizing that such structures are universal. With upper-grade classes, a simplified version of the finite-state automata concept (formalizing states and transitions) was introduced to foster progressive abstraction from concrete elements and prepare students for computational ideas.

The game fosters students to apply logical reasoning to forecast outcomes and to intuitively grasp concepts like state-dependent rules, with the additional support of a purpose-built software helping in verification activities and keyboard/mouse interaction skills testing.

Post-activity assessments revealed encouraging success rates with minimal teacher intervention.

Classroom observations enabled the calibration of a replicable (and documented) teaching protocol adaptable to different cognitive levels. The protocol's phased structure - exploration, guided practice, and open-ended problem-solving - proved crucial for supporting learning. Additionally, it provided operational insights for diagnosing and addressing prerequisite gaps, such as spatial reasoning difficulties in a small percentage of participants, mitigated through customized visual aids and direct mirror manipulation.

The results support the early introduction of informatics with a focus on structured understanding and problem-solving, challenging the notion that formal logic in languages is developmentally inappropriate for young learners. By integrating elementary informatics concepts into initial school curricula, this approach nurtures the development of meta-cognitive skills transferable to STEM disciplines.

Keywords: Informatics education, primary school, educational experimentation, languages, automata, storytelling.