Learning Transfer

Overview

Apply knowledge and skills from one domain to new contexts through systematic transfer strategies

Steps

Step 1: Extract deep structure from source

Analyze source knowledge to identify transferable elements:

1. Identify the underlying principles (not just procedures)
2. Articulate the relationships between elements
3. Distinguish essential structure from domain-specific details
4. Name the patterns at an abstract level
5. Ask: “What is this really about at its core?”

Abstraction questions:

• What problem does this solve?
• What are the key relationships/constraints?
• What conditions make this approach work?
• What would break this approach?
• What is the underlying logic?

Example:

• Concrete: “In chess, control the center to maximize piece mobility”
• Abstract: “In competitive systems, control key resources that enable the widest range of future options”

Step 2: Generate multiple source examples

Find varied examples of the same principle in action:

1. Identify other instances in the original domain
2. Find examples in related domains
3. Find examples in distant domains
4. Compare examples to identify common structure
5. Note what varies vs. what’s constant across examples

Why multiple examples matter:

• Single example: confuse surface features for deep structure
• Multiple similar: begin to see pattern
• Multiple varied: isolate true underlying structure

Comparison technique (Gick & Holyoak, 1983):

• Take two examples, ask: “What do they have in common?”
• The answer is more abstract than either example alone
• Repeat with more examples to refine abstraction

Step 3: Identify target domain structure

Analyze the target domain to find potential alignment:

1. Map the key elements of the target domain
2. Identify relationships between elements
3. Articulate the constraints and challenges
4. Note what’s similar to and different from source
5. Look for structural parallels, not surface matches

Analysis questions:

• What are the key entities in this domain?
• How do they relate to each other?
• What are the goals and constraints?
• What makes problems in this domain hard?
• What would success look like?

Step 4: Map source to target

Create explicit structural mapping between domains:

1. Align source elements to target elements
2. Map source relationships to target relationships
3. Check that mapping is systematic (not cherry-picked)
4. Identify disanalogies (where mapping breaks down)
5. Assess mapping quality

Mapping quality criteria:

• Structural consistency: Relations map to relations
• Systematicity: Connected system, not isolated correspondences
• Completeness: Important aspects covered
• Acknowledgment of limits: Disanalogies noted

Documentation format: Source Element → Target Element Source Relation → Target Relation [Note disanalogies and their implications]

Step 5: Generate transfer hypotheses

Derive specific predictions from the mapping:

1. Based on mapping, what should work in target domain?
2. What strategies/approaches should transfer?
3. What pitfalls should apply?
4. What should we expect to see if analogy holds?
5. What would indicate the analogy is misleading?

Hypothesis format: “If [source principle] transfers to [target domain], then we would expect [specific prediction] because [mapping justification]”

Hypothesis quality:

• Specific enough to test
• Derived from structural mapping (not surface)
• Includes expected observations
• Acknowledges uncertainty

Step 6: Adapt for target context

Modify source knowledge to fit target constraints:

1. Account for disanalogies identified in mapping
2. Adjust for domain-specific requirements
3. Translate terminology and concepts
4. Preserve essential structure while adapting surface
5. Create target-domain-appropriate implementation

Adaptation considerations:

• What’s different that requires adjustment?
• What domain conventions must be respected?
• How do constraints differ?
• What vocabulary does target domain use?
• What would “fluent” application look like?

Warning signs of poor adaptation:

• Forcing fit despite clear disanalogies
• Ignoring domain expertise in target
• Applying too literally without translation
• Overconfidence in transfer

Step 7: Test and refine transfer

Validate transfer through application:

1. Apply adapted knowledge in target domain
2. Compare outcomes to hypotheses
3. Note what transferred well vs. poorly
4. Refine understanding based on results
5. Update mental model of both domains

Testing approaches:

• Small-scale pilot application
• Thought experiment/simulation
• Consult target domain expert
• Compare to domain-native approaches

Interpretation:

• Success: Mapping was valid, refine and extend
• Partial: Some aspects transfer, others don’t - identify which
• Failure: Analogy misleading, reassess or abandon

Learning from transfer attempts:

• Success teaches about deep structure
• Failure teaches about important differences
• Both refine understanding of principles

When to Use

• Applying academic learning to professional practice
• Using expertise from one field in a new field
• Finding solutions from one domain applicable to another
• Building general problem-solving ability
• Learning something specifically for its transfer potential
• When solutions exist but aren’t being recognized as applicable
• Developing flexible expertise rather than narrow specialization
• Cross-functional work requiring integration of diverse knowledge

Verification

• Source knowledge is understood at principle level, not just procedure
• Multiple examples used to extract deep structure
• Target domain analyzed independently (not just assumed similar)
• Mapping is explicit, systematic, and acknowledges disanalogies
• Transfer hypotheses are specific and testable
• Adaptations account for domain differences
• Transfer is validated through application