How Preparation for Future Learning Works | Learn Today So You Can Learn Better Tomorrow

Quick Read: Transfer Is Not Only Carrying Old Knowledge Forward

John Bransford and Daniel Schwartz’s influential work on transfer argued that conventional assessment often defines transfer too narrowly: learn something in one context, then see whether the learner can apply it in another without additional help.

In their 1999 chapter Rethinking Transfer: A Simple Proposal With Multiple Implications, they proposed a complementary lens: preparation for future learning. Instead of asking only whether students can directly apply old knowledge, ask whether prior learning helps them learn more effectively from new resources and experiences.

Good learning should not only leave answers behind. It should leave the learner better equipped to extract value from the next answer, explanation, example, tool or experience.

The Closed-Book Transfer Test Is Only One Kind of Future

Imagine teaching a student a new mathematics concept today. Tomorrow the student receives an unfamiliar problem.

A classic transfer test might remove every resource and ask: Can the student solve it from existing knowledge?

That is important. Real independence matters.

But many real-world situations contain resources. Engineers consult standards. Doctors read new research. Programmers inspect documentation. Teachers learn from colleagues. Adults search, compare, ask, simulate and revise.

So another valuable question is: when the student is given a new worked example, explanation, data source or tool, can the student learn from it more efficiently and intelligently than someone who lacks the earlier preparation?

PFL measures readiness to learn in motion.

Performance Today Can Hide Readiness for Tomorrow

Two learners can obtain the same score and still be in different states.

Student A memorised a procedure and executes it accurately on familiar examples. Student B performs less smoothly but understands several contrasting cases, has noticed where intuitive methods fail and can explain the key variables.

On today’s narrow test, A may look better.

Tomorrow, the teacher introduces a new method. B may learn it faster because B has a richer problem representation and more differentiated prior knowledge.

PFL makes this possibility visible. Immediate efficiency and future learning potential are related, but they are not identical.

Why Prior Knowledge Matters Twice

Prior knowledge has an obvious role: it helps solve current problems.

It has a second role that is just as important: it determines what a learner can notice in new information.

An experienced reader notices an author’s argument structure because the relevant categories already exist. A strong mathematics student recognises that a new method solves a limitation encountered before. A science learner with a good causal model sees why a new experiment is diagnostic rather than merely procedural.

Knowledge changes the resolution of future learning.

The Problem of Teaching Answers Before the Learner Has a Question

Instruction often arrives before learners have encountered the problem that makes the instruction useful.

A formula appears before the learner has tried to represent the quantity it measures. An essay structure appears before the learner has experienced why an unstructured argument collapses. A scientific model appears before competing explanations have been considered.

The explanation can still be learned. But it may be encoded as another rule rather than as a solution to a recognised problem.

PFL-oriented instruction often creates intellectual preparation before the full explanation. Learners compare, invent, predict or attempt. They discover distinctions and limitations. Then formal instruction lands on a prepared structure.

Invention as Preparation for Future Learning

One well-known PFL design asks learners to invent a representation or procedure before receiving the canonical method.

The goal is usually not independent rediscovery of expert knowledge. The goal is to make learners explore the feature space.

Suppose students must compare several datasets. Before teaching a standard statistical measure, ask them to invent ways of representing “consistency” or “spread.” Their approaches may be incomplete. But in trying, students encounter the need to coordinate centre, deviation and distribution.

When the formal measure is introduced, learners can inspect what problem each component solves.

A 2023 study in Instructional Science describes invention-first instruction as a sequence of problem solving followed by explicit instruction, building on the PFL tradition of Schwartz, Bransford and colleagues.

Productive Failure Is a Route; PFL Is the Larger Outcome

Productive Failure and PFL are closely related but should not be collapsed into one owner.

Productive Failure is an instructional design: learners generate and explore before consolidation.

Preparation for Future Learning is a broader learning objective and assessment lens: does prior experience make the learner better at learning from what comes next?

Productive Failure can be used to build PFL, but PFL can also emerge from comparison, reflection, good foundational knowledge, resource-use skills, metacognition and other forms of instruction.

Series 0003 remains the dedicated owner for How Productive Failure Works.

The Four Capabilities Hidden Inside PFL

PFL is not one skill. It is a cluster of capacities that make future learning more productive.

1. Notice What Matters

The learner has enough structure to identify relevant features in a new explanation or problem.

2. Recognise What Is Missing

The learner can detect a gap instead of mistaking familiarity for completeness.

3. Use Resources Intelligently

The learner can search, select, interpret and integrate new information rather than simply copy it.

4. Update the Model

New information changes understanding. The learner does not merely accumulate another fact beside an unchanged misconception.

Dynamic Assessment: Test How Learners Learn With Help

Traditional assessment tries to standardise help away. Everyone gets the same conditions so performance can be compared.

PFL sometimes requires a different assessment design.

Give learners a novel problem. Then provide a new resource, worked example or mini-lesson. Measure not only the initial score but how much the learner improves after access to the resource.

This is a form of dynamic assessment. The dependent variable is not only what the student knew before help but how effectively the student converted help into capability.

Maria Mylopoulos and Daniel Schwartz argued in Medical Education that PFL is especially important in professions where practitioners must continually learn from new cases, evidence and resources.

A PFL Test Is Not an Open-Internet Test

Giving students Google access does not automatically measure PFL.

Resource-rich assessment needs a designed learning opportunity. The resource should contain information that can genuinely improve performance, and the task should reveal whether the learner can identify, interpret and integrate it.

A student who copies the first search result may have access without learning. A student who compares sources, maps the new principle onto the problem and updates a solution is demonstrating something closer to PFL.

Why Contrasting Cases Prepare Future Learning

One way to prepare learners is to show several cases that differ along meaningful dimensions.

If every example looks similar, students may encode surface features. Contrasting cases force attention toward the dimensions that vary and the relationships that remain.

Later instruction can name a structure the learner has already begun to perceive.

This is related to analogical encoding, but the ownership is different. Analogical encoding focuses on extracting shared relational structure through comparison. PFL focuses on whether such preparation improves learning from a subsequent opportunity.

See How Analogical Encoding Works.

PFL and Curiosity

A learner who has encountered a discrepancy has a question-shaped hole in the current model.

Why did two reasonable methods disagree? Why does this rule fail here? Which variable is missing? What would distinguish these explanations?

Instruction that arrives after such a discrepancy may receive more selective attention because the learner has a reason to care about the missing relation.

Curiosity is not guaranteed. But PFL-friendly designs often create information gaps with enough structure that the learner can recognise them.

PFL and Metacognition

Future learning depends partly on knowing what you do not know.

A learner who believes the model is complete will not search. A learner who recognises uncertainty can decide to seek evidence, ask a question or consult a resource.

This makes calibration important. Confidence should track the quality of the model closely enough that gaps trigger appropriate learning behaviour.

Metacognition does not replace knowledge. It tells the learner when the current knowledge is insufficient.

PFL and Search

Modern learners operate in an information-rich environment. The bottleneck is often not access to information but the ability to formulate useful searches and judge what comes back.

PFL therefore includes question quality.

A student who knows enough to ask, “What condition makes this theorem invalid?” has a better search state than a student who types “math answer.”

Better prior models produce better queries. Better queries produce better resources. Better resources improve the model.

That is a learning loop.

PFL and AI

AI makes PFL more important, not less.

A learner can obtain explanations, examples, critiques and simulations instantly. The educational question shifts from “Can you remember every answer?” toward “Can you recognise what you need, ask for useful information, evaluate the response and integrate it without losing independent judgement?”

A weak learner may use AI to bypass learning. A PFL-ready learner can use the same tool to extend a model, test an assumption and generate new practice.

The resource changed. The core capability did not: learn from what the environment can now provide.

PFL in Mathematics

Mathematics education often rewards clean execution of known procedures. PFL asks whether students are prepared to understand the next procedure when the old one reaches its limit.

Before introducing a formal method, students may compare cases, invent representations, estimate relationships or try incomplete strategies. The objective is not to delay instruction for ideological reasons. It is to prepare discriminations that make instruction meaningful.

After teaching, the PFL test gives a new resource and an unfamiliar problem. Can the learner extract the relevant idea and adapt?

Continue through the Mathematics Learning Hub.

PFL in English

Strong English learners do more than memorise model essays. They learn how to learn from texts.

Give two introductions and ask what each makes possible. Compare two paragraphs using the same evidence differently. Let students attempt an argument before showing a high-quality model. Ask them to revise after receiving a new rubric or reader response.

The PFL question is: does the learner know how to use the new model or feedback to improve the next piece?

Continue through the English Learning Hub.

PFL in Science

Science changes. New evidence arrives. Models are revised. A student trained only to reproduce accepted explanations has learned science content without fully learning the epistemic practice.

PFL-oriented science instruction gives learners opportunities to interpret data, encounter model limitations, seek missing information and update explanations.

The student should eventually be able to read a new source and ask: What evidence is new? Which part of my model changes? Which claim becomes weaker or stronger?

Continue through the Science Learning Hub.

PFL in Professional Training

Professional competence is especially vulnerable to obsolescence.

Standards change. Tools change. Evidence changes. New cases violate old routines.

Training cannot pre-teach every future case. It must also prepare professionals to detect when current routines are insufficient and learn efficiently from resources, colleagues and evidence.

This is why PFL has attracted attention in health professions education. The future expert is not simply the person with the largest static store of answers. It is the person whose knowledge makes subsequent learning faster, safer and more discriminating.

PFL and the School-to-World Handoff

School can create a dangerous illusion: that every problem will be preceded by the lesson that teaches exactly how to solve it.

The world is different.

Adults often meet the problem first. Then they must locate the lesson.

PFL prepares for that order. It builds learners who can enter an unfamiliar situation, identify what is missing and use available resources to close the gap.

Immediate Direct Instruction Still Matters

PFL should not be misread as an argument against explicit teaching.

Many domains contain conventions, safety rules, terminology and foundational procedures that should be taught directly. Novices cannot prepare for future learning if the present learning environment denies them essential structure.

The design question is sequencing.

When does a small amount of generation or comparison create useful preparation? When should the canonical explanation arrive immediately? Which components should be pre-trained first? Which later tasks should test resource use?

No single sequence owns every learning problem.

PFL and Pretraining

Pretraining and PFL may look opposite. Pretraining gives learners component knowledge before complexity; PFL often lets learners explore before full instruction.

They can work together.

Pretrain the minimum vocabulary or components required for meaningful exploration. Then give a problem that reveals deeper relationships. Then consolidate. Later assess whether the learner can use a new resource.

Series 0004 explains How Pretraining Works.

PFL and Cognitive Apprenticeship

Cognitive apprenticeship helps learners inherit expert ways of noticing, acting and monitoring. PFL asks whether those acquired structures make the learner better able to learn from later situations.

The relationship is powerful.

An apprentice who learns not only procedures but how experts seek information, compare alternatives and diagnose uncertainty becomes more prepared for future learning.

Series 0033 develops the apprenticeship mechanism in How Cognitive Apprenticeship Works.

The First Weak Link in PFL

If a learner fails to benefit from a new resource, diagnose the earliest broken operation.

• Did the learner fail to recognise the relevant feature?
• Did the learner fail to notice a gap?
• Was the search question too vague?
• Was the resource misunderstood?
• Did the learner retrieve an old rule that blocked updating?
• Was the new information understood but not integrated?
• Did the learner lack enough foundational knowledge to use the resource at all?

“Cannot learn independently” is too large a diagnosis. PFL improves when the failing sub-process becomes visible.

A PFL Lesson Architecture

• Establish the floor: ensure enough prior knowledge exists to make exploration meaningful.
• Present a structured challenge: use a problem with dimensions the later instruction will organise.
• Generate: let learners produce representations, predictions or candidate methods.
• Compare: surface similarities, differences and failures.
• Consolidate: provide explicit instruction that connects directly to the explored features.
• Retrieve: stabilise the new model.
• Delay: allow time and interference.
• Provide a new resource: a worked example, data source, explanation or tool.
• Assess learning from the resource: measure how much capability changes.
• Reflect: ask what the learner noticed, sought and updated.

A Home-Study Version

Students can use PFL without redesigning the entire curriculum.

Before opening the worked solution, spend five minutes trying the problem. Before reading the chapter explanation, compare two examples and predict the rule. Before watching a tutorial, write what you currently think the mechanism is.

Then consume the resource actively. Ask: What does this explain that my attempt could not? Which distinction was missing? What should I update?

The attempt is valuable only if it changes how the resource is processed.

A Parent Version

When a child asks for help, parents often face a false choice between giving the answer and refusing help.

A PFL response can be: “Show me what you tried first. Now let us look at one example together. What in this example changes your plan?”

The parent is not measuring whether the child could solve everything alone. The parent is teaching the child how to learn from support.

A Tutor Version

Tutors can create PFL by preserving a small amount of productive uncertainty before explanation.

Ask students to compare cases, commit to a prediction or build a first representation. Then teach. After teaching, explicitly revisit the initial attempt.

The tutor’s explanation should answer a problem that now exists in the learner’s model.

The Danger of Measuring Only Immediate Efficiency

School systems naturally prefer metrics that are easy to collect. Accuracy today is easier to measure than learning potential tomorrow.

But optimisation against immediate scores can produce brittle learning. Teachers may remove every difficulty, supply every representation and rehearse exact formats until practice performance is excellent.

If students then fail to learn from a changed situation, the system trained performance without adaptability.

PFL adds a longer horizon to instructional design.

The Danger of Romanticising Struggle

The opposite mistake is equally serious.

Not every struggle prepares future learning. Confusion can remain confusion. Failed search can consume time. Learners can generate misconceptions and reinforce them.

Preparation must be designed around a later learning opportunity. If the initial challenge does not activate relevant structures, expose useful distinctions or create productive questions, it has little reason to improve future learning.

What Should Be Measured?

A PFL assessment can track several things.

• initial performance before new information;
• quality of questions or resource selection;
• ability to extract relevant information;
• degree of improvement after the resource;
• quality of the updated explanation;
• transfer to another case;
• and efficiency of subsequent learning.

The strongest measure is not always a single score. It may be the trajectory from first attempt to informed second attempt.

PFL and Lifelong Learning

Lifelong learning is often described as motivation: keep learning throughout life.

Motivation matters, but desire without learning architecture is not enough.

Adults need models that help them identify what has changed, locate relevant knowledge, evaluate sources, integrate new evidence and revise routines without discarding everything already known.

PFL is therefore one of the deep competencies behind lifelong learning. It asks whether education has created a person who can continue updating after formal teaching ends.

The Deep Principle: The Best Lesson Changes the Value of the Next Lesson

Learning has compound effects.

A strong concept today makes tomorrow’s explanation easier to understand. A useful distinction today makes tomorrow’s example more informative. A good question today makes tomorrow’s search more precise.

This means educational value is not fully visible at the moment of acquisition.

Some knowledge produces direct performance.

Some knowledge increases the learner’s capacity to acquire more knowledge.

The second effect is quieter, but over years it may be larger.

Research and Further Reading

• Bransford & Schwartz — Rethinking Transfer: A Simple Proposal With Multiple Implications
• Mylopoulos & Schwartz — Preparation for Future Learning: A Missing Competency in Health Professions Education?
• Effectiveness of Invention Tasks and Explicit Instruction in Preparing Learners for Learning
• How Productive Failure Works
• Study & Learning Methods Hub

---

eduKateSG Learning Node Series · 0034 of the continuing series. Previous: 0033 — How Cognitive Apprenticeship Works. Continue through the Study & Learning Methods Hub.