Purpose of the Unit

When learners encounter a question, they usually rush straight towards an answer. Mindgle teaches them to pause at the question and examine it first.

A question is not a narrow corridor with only one route; it is a room with several intellectual doorways.

One doorway opens the meaning of the words. Another exposes assumptions. Another reveals exceptions, missing voices, competing perspectives, hidden causes, or future consequences. The learner does not have to enter through every doorway, but must learn to choose the doorway—or combination of doorways—that will produce the most meaningful inquiry.

To make the architecture clear, let us use one raw question throughout:

Does AI improve learning?

Keyword means the word carrying the greatest conceptual weight in a question.

Archaeology means excavating beneath what is visible to discover earlier layers. Keyword Archaeology therefore means digging beneath the present meaning of a word to uncover its origin, development, and changing use.

The learner identifies the most important words in the question and investigates:

• their etymology or linguistic origin;
• their original meaning;
• how their meanings have changed;
• how different disciplines define them;
• whether people are using the same word to mean different things.

In “Does AI improve learning?”, the words intelligence, improve, and learning need investigation.

The word intelligence comes from the Latin intelligere, meaning “to understand,” “to discern,” or “to choose between.” This immediately creates a deeper question: if intelligence involves genuine understanding and discernment, in what sense is artificial intelligence actually intelligent?

The word learning is equally important. Does learning mean remembering information, gaining marks, completing a task, changing one’s understanding, acquiring a skill, or becoming capable of acting independently?

What this doorway reveals

Familiar words often create an illusion of shared understanding. Two people may agree that AI “improves learning” while one means higher examination scores and the other means deeper conceptual understanding. Unless the key words are clarified, the inquiry may rest on conceptual confusion.

Doorway questions

• Which word controls the meaning of this question?
• Where did this word originate?
• What did it originally mean?
• Has its meaning changed over time?
• Do different disciplines define it differently?
• Are we using one word for several different phenomena?

Possible output

“What should count as genuine improvement in learning when students use  AI?”

Ontology is the study of what exists and what kind of reality something possesses.

Unpacking means opening a broad or compressed idea to see what is actually contained within it.

Ontological Unpacking asks the learner to move beyond the word and investigate the reality to which the word supposedly refers.

For example, we frequently use the word learning, but what exactly is learning? Is it:

• an internal change in understanding;
• an observable change in behaviour;
• the ability to remember information;
• the ability to transfer knowledge to a new situation;
• improved performance in an assessment;
• or a combination of these?

A learner may produce an impressive AI-assisted assignment without experiencing any meaningful cognitive change. The product has improved, but has learning occurred? The Reality Doorway separates the visible product from the invisible process.

It also asks whether a category exists naturally or has been constructed by people. “Academic ability,” “intelligence,” “success,” and even “failure” may not be single, fixed realities. They may be categories shaped by culture, institutions, assessment systems, and historical circumstances.

What this doorway reveals

The existence of a word does not prove the existence of one clear and measurable reality behind it. Before investigating whether something changes, causes, or improves something else, learners must establish what that “something” actually is.

Doorway questions

• What exactly is the phenomenon being discussed?
• What would count as evidence that it exists?
• Is it directly observable or only inferred?
• Is it a natural reality, a social construction, or a combination of both?
• Where does this phenomenon begin and end?
• Are we treating several different realities as if they were one?

Possible output

“To what extent should learning be understood as improved performance, internal cognitive change, or the independent transfer of understanding?”

A premise is an idea that must be accepted as true for a question or argument to make sense.

An assumption is something treated as true without first being adequately examined or demonstrated.

Premise Excavation means uncovering what the question has quietly accepted before the inquiry has even begun.

“Does AI improve learning?” may be assuming that:

• AI use can be clearly defined and measured;
• learning can be measured accurately;
• improvement has the same meaning for every learner;
• AI is the factor producing the improvement;
• better work indicates better understanding;
• the effects of AI are similar across subjects and age groups.

These assumptions may be reasonable, partially true, false, or impossible to establish without further investigation. The learner’s task is not to reject every assumption but to make each one visible and testable.

Consider the question, “Why are students becoming less attentive?” It already assumes that attention is declining. It may also assume that students themselves are responsible. If neither assumption is examined, the research may simply collect evidence that confirms the question’s original judgement.

What this doorway reveals

A biased premise can quietly predetermine the conclusion. Premise Excavation prevents learners from conducting an entire investigation inside an assumption they never noticed.

Doorway questions

• What is this question already treating as true?
• What must I believe for this question to make sense?
• Has that belief been demonstrated?
• Does the wording assign responsibility before evidence is collected?
• Would the question collapse if its central assumption were false?
• How can I rewrite it so the assumption becomes investigable?

Possible output

“Under what conditions, if any, is the use of AI associated with measurable changes in students’ learning?”

The words “if any” are powerful because they leave open the possibility that the presumed effect may not exist.

A concept is a broad mental category used to organise and understand reality.

Fracturing means splitting something that appears whole into its constituent dimensions.

Large concepts are often too compressed to investigate. Concept Fracturing breaks the central idea into its hidden components.

In the question “Does AI improve learning?”, the word improve might mean:

• increasing examination scores;
• improving the quality of written work;
• accelerating task completion;
• strengthening memory;
• deepening conceptual understanding;
• developing creativity;
• increasing confidence;
• improving problem-solving;
• supporting independent learning;
• enabling transfer to unfamiliar contexts.

These outcomes are not identical. AI might improve one while weakening another. It might increase the quality of a submitted response while reducing independent thinking. Without fracturing the concept, the learner may produce a conclusion that is technically true but intellectually misleading.

What this doorway reveals

A broad concept may conceal several variables that behave differently. Concept Fracturing converts a vague curiosity into a precise and researchable inquiry.

Doorway questions

• What different meanings are compressed inside this word?
• What are the dimensions of this concept?
• Can one dimension increase while another decreases?
• Which dimension matters most for this inquiry?
• Which dimensions can be observed or measured?
• Am I making a general claim about something that is internally diverse?

Possible output

“How does generative AI influence students’ conceptual understanding, task performance, retention, and intellectual independence?”

An anomaly is an observation, case, or result that does not fit an expected pattern or accepted explanation.

Anomaly Hunting means deliberately searching for cases that challenge what appears normally or generally true.

Suppose evidence shows that students using AI produce better assignments. Instead of immediately concluding that AI improves learning, the learner searches for exceptions:

• students whose work improves but whose understanding declines;
• students who become more independent through AI;
• students who become dependent upon it;
• subjects in which AI supports learning;
• subjects in which it creates misconceptions;
• learners who benefit because of language barriers or learning differences;
• learners whose confidence grows but whose accuracy does not.

The exception is not an inconvenient detail to be removed. It may reveal that the original explanation is incomplete or that the effect depends on particular conditions.

What this doorway reveals

A theory is often most informative at the place where it stops working. Anomalies expose hidden variables, boundary conditions and alternative explanations.

Doorway questions

• Where does the expected pattern fail?
• Who does not fit the general conclusion?
• Under what conditions does the opposite occur?
• Is this truly an exception, or evidence that the original pattern was oversimplified?
• What hidden factor might explain the anomaly?
• How must the claim change to accommodate this case?

Possible output

“Why does AI strengthen independent learning for some students while producing cognitive dependence in others?”

Inversion means turning the direction or order of a relationship around.

Inquiry Inversion asks learners to reverse the question rather than accepting its original direction.

The conventional question asks:

How does AI change human learning?

The reversed question asks:

How do human learning habits influence the way AI is used?

A second reversal could be:

Instead of asking whether AI makes students dependent, how might students’ existing dependence on quick answers shape the development and use of AI?

This helps the learner recognise that many relationships are not one-directional. AI may influence learners, but learner behaviour, school assessment, commercial priorities and cultural expectations also influence how AI develops and functions.

The relationship may therefore be reciprocal:

AI changes learning behaviour, while learning behaviour changes how AI is designed and used.

What this doorway reveals

Reversal exposes feedback loops, reciprocal relationships and circular causation. It prevents the learner from automatically treating one factor as the cause and the other as the passive effect.

Doorway questions

• What happens if I reverse the cause and effect?
• Could the supposed outcome also influence the supposed cause?
• Is the relationship linear, reciprocal, or circular?
• Who or what appears passive in the original question?
• What changes when I give that factor agency?
• Could both variables be influencing each other?

Possible output

“How do students’ existing learning habits shape their use of AI, and how does repeated AI use subsequently reshape those habits?”

In art, negative space is the space around or between visible objects. It may appear empty, but it influences how the entire image is understood.

In inquiry, negative space refers to what is absent:

• voices not heard;
• groups not represented;
• evidence not collected;
• experiences not recorded;
• consequences not discussed;
• questions not asked;
• knowledge not recognised as legitimate.

The learner examines a discussion not only for what it contains but also for what it leaves outside.

Research claiming that AI improves learning may be based on students who have reliable internet access, English-language fluency, modern devices and experienced teachers. The inquiry may remain silent about rural learners, low-connectivity communities, learners with disabilities, teachers with limited training, or students whose first language is not well represented in AI systems.

Silence may also exist in the evidence. Examination scores may be measured while curiosity, agency, anxiety, originality and long-term retention remain unmeasured.

What this doorway reveals

Absence is not always accidental or neutral. What research leaves out can shape its conclusions as powerfully as what it includes.

Doorway questions

• Who is speaking?
• Who is being spoken about but not heard?
• Which population is absent from the evidence?
• What outcome has not been measured?
• Whose knowledge has not been treated as credible?
• Who benefits from this silence?
• How might the conclusion change if the missing voice were included?

Possible output

“How do unequal access, language, disability and teacher preparedness shape students’ experiences of AI-supported learning?”

A perspective is a way of interpreting something from a particular position, experience, discipline, culture or set of interests.

Perspective Rotation means deliberately turning the question through different lenses instead of allowing the first viewpoint to represent the whole truth.

Consider AI-supported learning through different lenses:

None of these lenses alone provides the complete reality. Each makes certain aspects visible while placing others outside its field of attention.

What this doorway reveals

Knowledge is shaped partly by the position from which reality is observed. Perspective Rotation does not mean that every perspective is equally accurate; it means that each must be examined, compared and supported with appropriate evidence.

Doorway questions

• Who would see this issue differently?
• How would another discipline frame the question?
• What does this perspective make visible?
• What does it conceal?
• What interests or experiences shape this interpretation?
• Where do the perspectives agree, conflict or complement one another?
• Which perspective is best supported by evidence?

Possible output

“How do learners, teachers, neuroscientists and ethicists differ in their understanding of what AI-supported learning improves or diminishes?”

Magnitude refers to size, level or extent.

Magnitude Shifting means examining the same phenomenon at different levels of reality.

Something beneficial at one scale may become harmful at another. AI may help one student overcome a language barrier, but widespread uncritical use may weaken a school’s culture of independent inquiry. A short-term benefit may also create a long-term cost.

Magnitude Shifting can also involve time:

• immediate versus delayed effects;
• short-term performance versus long-term retention;
• present convenience versus future intellectual independence.

What this doorway reveals

A conclusion may be accurate at one scale but misleading at another. The learner must specify whose reality, which level and what period the claim describes.

Doorway questions

• At what level am I examining the issue?
• Does the explanation remain valid when the scale changes?
• Could an individual benefit create a collective disadvantage?
• Could a short-term gain produce a long-term loss?
• Is an individual behaviour being caused by an institutional or systemic condition?
• At which scale would intervention be most effective?

Possible output

“How do the individual benefits of AI-assisted learning compare with its long-term effects on classroom culture and learner independence?”

A counterfactual is an imagined situation in which one feature of reality is changed, removed or reversed.

Simulation means mentally modelling what might happen under that altered condition.

Counterfactual Simulation invites learners to construct an “impossible” or alternative world and observe what changes.

For example:

If AI disappeared tomorrow, which learning improvements would remain?

If the student could still explain, apply and transfer the learning independently, AI may have supported genuine learning. If the quality disappeared as soon as the tool was removed, AI may have improved only the product.

Other counterfactuals could include:

• What if AI could ask questions but never provide answers?
• What if students could use AI during learning but not during assessment?
• What if every AI-generated statement had to be independently verified?
• What if schools assessed thinking processes rather than final products?
• What if AI access were available only to wealthy learners?

Counterfactual thinking does not prove causation by itself, but it helps learners test causal explanations and distinguish necessary conditions from contributing conditions.

What this doorway reveals

When a factor is mentally removed and the outcome remains unchanged, that factor may not be the true cause. Counterfactuals expose causal dependence, alternative possibilities and unintended consequences.

Doorway questions

• What would happen if this factor disappeared?
• What would remain unchanged?
• What if the opposite condition existed?
• Is this factor necessary, sufficient, contributory or merely associated?
• Which unintended consequences would appear?
• What does this imagined world reveal about the real one?

Possible output

“If AI support were withdrawn, which forms of understanding and capability would students retain independently?”