Formation Through Practice

Purpose of this document

Vocational learning in the AI era. What’s valuable to learn, and how to teach it?

Use this as a working synthesis, not as settled doctrine. Several ideas below are hypotheses to examine, refine, and test.

1. Starting point: “just doing” is no longer enough

Just doing an assignment isn’t evidence of learning. It isn’t likely to add human value either. When the required outcome is well-defined, an automation could “do” it.

Then, what value do you add? How do you train for it? How do you evaluate, to ensure that you trained well?

2. Traditional approaches in times of AI agents

Historically, writing something (prose, code) was evidence that the writer can apply that knowledge. AI-agents weaken this link: the outcome can become ‘good enough’ before its author understands it.

Then is “understanding” the key?

LLMs can recall facts, apply familiar patterns, and sometimes reason from first principles. Therefore, simply teaching people to remember facts or imitate expert application is insufficient.

Our “understanding” needs to reach deeper. But then what’s “deeper”?

3. From knowledge gaps to modes of inquiry

Between what, how, and why

You learn a tool by asking what it does, and how to use it. Asking “why did someone make it?” can tell you whether it would be effective in your situation. This is “design intent”

Inquiry modes - a practical inventory

• Design intent: Why was it designed this way? Will it sustain in my context?
• Mechanistic understanding: How does it work? What does it depend on? What impact does it have?
• Epistemic maturity: How confident should I be, and what would change my mind? What assumptions am I making?
• Empathy: How do users flourish with the product?

Example shift in inquiry:

• From: “How do we get patients to comply with their treatment?”
• To: “What makes this treatment hard to follow in their lives?”

The shift: From jumping to fix non-compliance to investigating the patient’s reality.

4. Structural transfer and fundamental principles

A durable capability is the ability to recognize fundamental principles across different domains and transfer understanding without assuming that surface similarities imply identical solutions.

A fundamental principle is not merely a vaguer or broader statement. It should reveal causal structure, constraints, invariants, or generative relationships.

5. Intent, attention, and framing

The abstractions that people notice and transfer depend partly on what they are trying to achieve. The same engineering choice—such as keeping a design simple—can arise from very different motives:

• Learning what users actually need before committing;
• Fear of unfamiliar infrastructure;
• A short deadline;
• A desire for recognition;
• Concern for maintainability or operational cost.

Therefore, the outward action alone does not reveal its generator, or the ability to apply it when required.

6. Self-discovery as part of the work

Internalizing something lets it endure beyond the present. It’s not a separate retrospective activity. The artifact and the ongoing work can expose:

• Gaps in understanding;
• Assumptions about the world;
• Habitual ways of framing problems;
• Fears, attachments, or avoidance;
• What one cares about;
• The purposes one wants to serve.

The answer might be a more generative model, a recognized assumption, a new way of seeing.

7. A working three-part orientation for every activity

In every activity, keep three concerns alive:

1. Reality: Expand my interpretation

What is reality teaching me that my current model cannot yet explain?

Acknowledge that:

• Understanding is always partial;
• Consequences may be unforeseen;
• Reality can contradict elegant designs and confident explanations;
• Failures, invariants, mechanisms, design intent, and other perspectives can refine the model.

2. Building well-wishers: Nurture people to flourish through trust

Who can work or live better because of this, and how do I build trust?

This is broader than satisfying immediate user requests. It involves understanding people’s actual circumstances, enabling their agency, and caring about consequences. Empathy without discernment can become people-pleasing; concern for others must remain answerable to reality and purpose.

3. Self-formation: Seek enduring, transferable understanding

What do I take back for myself from this experience?

A journey of “uplifting”:

• Move from incidents and local solutions toward mechanisms, constraints, and generative principles;
• Seek explanations that transfer to other domains;
• Notice and relinquish unnecessary assumptions or fixations;
• Carry forward a refined way of seeing, not just another recipe.

8. Integrity as alignment

Integrity is alignment among:

• What reality permits or requires;
• What nurtures people;
• What the practitioner intends;
• What the artifact actually does.

9. The project

Two connected but distinct projects.

Project A: An AI-assisted vocational-learning tool

Question: How can an AI-assisted learning experience map a learner’s path of discovery, without becoming an interrogation?

This aids continuous evaluation by the learners themselves. It gives direction towards the next step.

Project B: What vocational education should cultivate in the AI era

Question: What should education aim to form when AI can increasingly produce artifacts and apply established expertise?

Likely topics:

• structural transfer;
• problem framing and cognitive orchestration;
• intent, attention, and self-awareness;
• integrity/alignment among reality, people, and self.

10. Implementation challenges

1. Fundamental principles: How can learners distinguish generative principles from vague generalizations? Test explanatory power, prediction, transfer, and boundary conditions.
2. Empathy/self-awareness reciprocity: Can learners cultivate this by collaborating on shared artifacts?
3. Integrity/alignment: Is alignment a condition, a process, an emergent property, or an ideal? How can learners recognize it, without turning it into a checklist?
4. Fixation and insight: When does relinquishing an assumption reveal a better representation, and when is it merely abandoning a valid constraint?
5. Anti-sloppiness: Reality-orientation, care for others, and self-formation may not by themselves ensure careful execution. How should professional standards, verification, and respect for the artifact fit in, especially as execution becomes automated?
6. Assessment: How can the learning tool detect actual development in a learner’s habits of inquiry rather than reward polished self-reports or compliance with prompts?

References

These are starting points for research, not a claim that every author endorses the complete framework:

• John Dewey: reflective thought, inquiry, experience, education and growth.
• Donald Schön: reflection-in-action and reflection-on-action in professional practice.
• David Kolb: experiential learning cycle.
• Jack Mezirow: transformative learning and revision of frames of reference.
• Jean Lave and Etienne Wenger: situated learning, communities of practice, identity through participation.
• Aristotle: virtue, habituation, practical wisdom (phronesis).
• Alasdair MacIntyre: practices, internal goods, standards of excellence.
• Chris Argyris: espoused theory vs. theory-in-use; double-loop learning.
• Peter Senge: personal mastery and the tension between vision and current reality.
• Gregory Bateson: levels of learning.
• Piaget: assimilation and accommodation.
• Constructivism, analogical reasoning, schema abstraction, conceptual change, and transfer of learning: how reuse & reorganization happen with prior understanding.
• Robert Selman: perspective-taking development.
• Carl Rogers: empathic understanding and respect for the other person’s subjective experience.
• Role-taking literature: links among perspective-taking, empathy, and self-understanding.
• Richard Sennett, The Craftsman; Matthew Crawford, Shop Class as Soulcraft; David Pye: workmanship, craft, and the relation between doing and judgment.