Skip to content

Understanding must be cultivated through exploration, not described or told

Principle · All principles
Evidence · 22 claims (13 for, 5 mixed, 4 against) · 26 studies (7 causal, 5 review, 4 quant-synthesis, 4 design, 3 qualitative, 3 theoretical), q1–q4 · 5 of 26 report an effect size · 17 claims rest on one study

Conditional relationship

This page owns how a teacher guides while learners explore a new idea: by clues, questions and critique that work on the learners' own attempts, or by stating the principle and explaining it. For a learner meeting a concept or principle (in the source studio, design principles such as balance, rhythm or hierarchy; elsewhere, an explanatory idea in a discussion) who already knows enough to make and talk about attempts, the proposed relationship is: a period of exploration in which the teacher sets a problem with a clue, asks questions that sharpen the learner's own partial idea, and leads a critique of everyone's work, followed by naming the principle with the learners' own works as the cases, is expected to produce better conceptual-understanding and transfer to new cases (near-transfer) than either being told the principle first and then practising it, or exploring with no guidance. It is not expected to help with facts to be remembered, with procedures the learner has never seen, or with domain-general skills such as designing a fair test.

No claim in the wiki tests this page's relationship as such: none compares indirect, clue-and-question guidance with direct explanation of the same principle in a studio or discussion, on an outcome measured afterwards. What is recorded is (a) qualitative episodes in which a teacher's guiding question let students integrate ideas while a teacher's interjected explanation shut their talk down, (b) one small tutoring study in which prompting without explaining taught as well as explaining, and (c) meta-analyses showing that exploration with guidance beats other instruction, unguided exploration loses to explicit instruction, and problem solving before instruction helps conceptual learning only when instruction follows. That last finding revises the earlier page: the evidence recorded here supports exploring before being told and being guided by questions, not never being told. The principle is still named, explicitly, after the exploration and from the learners' work.

The relationship is conditional on the learner's prior knowledge of the target (guidance matters most for novice learners and its advantage recedes with high prior knowledge), on the kind of knowledge (concepts and principles, not domain-general skills or facts), on age (problem-first sequences reversed for second to fifth graders), on whether anyone can ask the questions during the exploration, and on the horizon.

How the converted siblings sit beside this page: Productive Failure owns whether a new concept is attempted or taught first; this page assumes the attempt comes first where that page's conditions hold and adds how the teacher talks during it. Inquiry-Based Learning owns guided against unguided inquiry. Direct Instruction owns explanation, modelling and guided practice, which is the right default where this page's conditions fail. Error Analysis owns using errors as the object of study, which this page's critique step draws on. Peer Discussion owns discussion after an individual commitment. Technology should augment, not supplant, learning processes applies the same division to a tool that could give the answer.

Default design, while the relationship is untested

The earlier page's guidance (an architecture basic-design studio running six hours a week over a 14-week term), kept as a concrete default and corrected where the evidence disagrees. Each step is a design proposal; its evidence status is stated beside it.

  1. Open with a problem and a clue, not the principle. Each week, give a clue drawn from another discipline (a musical rhythm, a biological pattern, a typographic grid) and a design problem it bears on; learners first analyse the clue (what is it made of, how are its parts organised) and then make their own composition. In a six-hour studio: about 30 minutes on the clue, in pairs. Untested proposal, from the earlier page and its clues strategy; the one observation recorded is that students later used the clues they found when explaining their designs (Students exemplified discovered clues when explaining their designs orally [+W]).
  2. Give real time to make attempts. About two to two and a half hours of the six-hour session for making, with at least two versions of the composition. Exploration with guidance is supported in general (Enhanced discovery learning produces greater learning than other forms of instruction [+M]); the minutes are an untested proposal.
  3. Guide by questions on the learner's own idea, not by explanation. Circulate during making; wait until a learner has a position, then ask one specific question that sharpens something the learner has already half-said ("what is holding the left side down?"), and confirm the right term when a learner reaches for it, without adding the explanation. Do not interrupt a group's reasoning to state the principle. Basis: in recorded small-group episodes a teacher's guiding question let students' own partial idea become the shared concept, while a teacher who interjected disciplinary content was resisted and the talk stopped being dialogic (Subtle teacher guidance, not imposition, enables interdisciplinary integration in small-group deliberation [+W]); in one tutoring study, prompting without explaining taught as well as explaining (Human tutoring has a medium effect over no tutoring, and in one small study students learned as well when tutors only prompted them as when tutors also explained and gave feedback [+M]). Avoid closing remarks ("fine", "OK, next") that end a learner's line of thought before it is finished (A 1970 teacher self-analysis programme asserts, without data, that teacher responses including verbal rewards can close off pupil thinking [+W]).
  4. Critique every work, strengths and weaknesses together. Pin up all compositions; for each, the class says what works and what does not, against the week's principle, without sorting works into good and bad and without a single "right" answer. Treat a weak composition as data about the principle. About 90 minutes for a studio of 20 to 25. Untested as a studio routine, from the earlier page; comparing correct with incorrect examples is supported for concepts in mathematics (Erroneous examples improve conceptual understanding by forcing comparison with correct models. [+M], carried here from decimal magnitude with children to design critique with adults). See Error Analysis.
  5. Each learner explains their own work aloud. Two or three minutes each during the critique: what the clue was, what they did with it, and why. Untested; from the earlier page; the observation in step 1 is the only record ([+W] above). See Self-Explanation.
  6. Then name the principle, from their work. In the last 20 to 30 minutes, state the principle explicitly and its usual name, pointing to the pinned-up works that show it and to those that break it. This replaces the earlier page's "avoid direct narration": in the evidence recorded here, problem solving before instruction helps conceptual learning when instruction follows and builds on the learners' solutions (Productive Failure Improves Conceptual Learning [~S]; PS-I fosters learning only if specific design features (contrasting cases or building instruction on student solutions) are implemented [~M]), and exploration with no guidance learns less than explicit instruction (Explicit instruction produces greater learning than unassisted discovery learning under most conditions [-M]).
  7. Check on a new case, without the clue. End the week, or open the next, with a short task the clue does not cover: a new composition to judge or adjust, explained in a sentence or a sketch. Untested proposal; it shows whether the principle transfers beyond the clue that carried it.
  8. Hand over the questioning across the term. Weeks 1 to 4: a strong clue close to the principle, the teacher asks most questions, and names the principle at the end of every session. Weeks 5 to 9: looser clues; learners lead the critique with a short question card ("what holds this together? what breaks it?") and the teacher adds one question per work. Weeks 10 to 14: learners choose or bring their own clues, critique one another's work, and name the principle themselves; the teacher corrects only what is wrong. A learner moves to the looser stage when they can name and use the week's principle on a new case unaided in two consecutive weeks. The arc and its threshold are untested proposals; that guidance matters less as prior knowledge grows is a review's statement (The advantage of guidance begins to recede only when learners have sufficiently high prior knowledge to provide internal guidance [~M]).
  9. Tell first when the conditions fail. If the target is a procedure learners have never seen (a drawing convention, software operation), a domain-general skill, or the learners are young children, explain and model first and let them explore afterwards (For domain-general skills, one meta-analysis finds a negligible effect (g = −0.17) in favour of I-PS over PS-I [-M]; the age reversal in the productive-failure claim above). See Direct Instruction.

When a condition is not met: if nobody can ask questions during the exploration (self-paced online work), build the questions into the task as prompts, and show contrasting example works with a short explanation after the learner's attempt; if the session is short, keep the order (attempt, critique, name, check) and shrink each part.

Fitting the design to a situation

The three facts that most change the decision:

  • What the learner already knows about this target, and what kind of target it is. Clue-led exploration needs learners who can make attempts and talk about them, and a concept or principle to find; a procedure never seen, a domain-general skill or a set of facts is better told first. If unstated, ask: can learners already produce something, however rough, in this domain, and is the target an idea to understand or a step sequence to follow?
  • Who can ask the questions during exploration, and how often. The evidence for guiding by questions comes from teachers sitting with small groups and from one-to-one tutoring; with one teacher for many groups, or none, the questions must be prepared in advance or built into the task. If unstated, ask: during making, who will see each learner's or group's attempt and talk with them, and for how long?
  • Whether the principle will be named after the exploration, and what outcome counts. Exploring before being told helps conceptual understanding when instruction follows; it does not help procedural fluency. If unstated, ask: what must the learner be able to do at the end (judge a new case, explain why, produce work), and when?
If the brief says… Then change… Basis
Learners are novice on a procedure or technique they have never seen (a drawing convention, software steps) Demonstrate and explain the procedure first; keep clue-led exploration for the principle the procedure serves Explicit instruction produces greater learning than unassisted discovery learning under most conditions [-M] (meta-analysis across domains; not studio-specific)
Learners are advanced on the target, or already name and use the principle Fewer, looser clues; learners set their own problems and lead the critique; the teacher only asks The advantage of guidance begins to recede only when learners have sufficiently high prior knowledge to provide internal guidance [~M] (a review's statement; gives no crossover point)
Learners are child (about 7 to 11) Name and show the idea first, then a short guided exploration with the teacher at the table; keep critique to two or three works Productive Failure Improves Conceptual Learning [~S] (problem-first reversed for second to fifth graders); Age moderates discovery-learning effects: the explicit-instruction advantage over unassisted discovery is larger for adolescents than adults, and the enhanced-discovery benefit larger for adults than children [~M]
Learners are adolescent Keep exploration guided throughout (prepared questions, a teacher check-in per group); never leave a group to discover alone Age moderates discovery-learning effects: the explicit-instruction advantage over unassisted discovery is larger for adolescents than adults, and the enhanced-discovery benefit larger for adults than children [~M]
Learners are adult in a work setting (adult-workplace) Start from a case from their own work as the problem; use their experience as material the analysis is built on, not a warm-up; name the principle with their cases Untested for adults at work; carried from a high-school English unit (Positioning personal experience as an epistemic resource supports expansion from inward to outward critical orientations [+W])
Goal is information to remember (facts, terms, rules to recall) Tell it, clearly and briefly; keep exploration for the parts that need understanding Discussion promotes more active thinking than lecture, though lecture is no worse for transmitting information [~M] (college teaching, older reviews)
Goal is a domain-general skill (designing a fair test, a general reasoning strategy) Teach the strategy first, then practise it on varied problems For domain-general skills, one meta-analysis finds a negligible effect (g = −0.17) in favour of I-PS over PS-I [-M]
One teacher with a large group or many groups Groups explore; the teacher circulates with three or four prepared guiding questions per week, joins each group once, and critiques a sample of six to eight works with the whole room Untested at scale; the recorded episodes show the unprepared teacher interjecting explanation was the more frequent pattern (Subtle teacher guidance, not imposition, enables interdisciplinary integration in small-group deliberation [~W])
Facilitator time is scarce (one tutor dropping in, volunteers) Put the guiding questions on a card the group works through; the tutor's one visit asks a single question on the group's own partial idea and checks the closing naming of the principle Untested proposal
Single session only, or a fixed short time (for example 50 minutes) 5 minutes clue and problem; 15 minutes making; 15 minutes critique of four or five works; 10 minutes naming the principle from those works; 5 minutes on one new case. Do not claim lasting understanding Untested proposal
Some learners remote, others in the room; or learners on phones only Learners photograph or upload each version to one shared board; critique on the shared screen with remote learners speaking first; questions in the chat when voice is not possible Untested proposal
online-self-paced, no teacher present Build the questions into the task as prompts after each attempt; then show contrasting example works (one that shows the principle, one that breaks it) and a short explanation that refers back to the learner's attempt Untested here; carried from mathematics: Erroneous examples improve conceptual understanding by forcing comparison with correct models. [+M]; PS-I fosters learning only if specific design features (contrasting cases or building instruction on student solutions) are implemented [~M]
Learners have low confidence, low literacy, or find public critique hard Critique the work, not the maker; let works be shown without names; allow a learner to explain by pointing and sketching; open each critique with what works Untested proposal; see impersonal critique
The material is private or personal (own experiences, own life as the problem) The learner decides what is shown to the group and what stays with them; critique only what is shared Untested proposal
The brief names its own critique routine, rubric or criteria (a jury format, a programme rubric, its own categories) Keep the routine and its categories as given; change only the teacher's move during making (a question instead of an explanation) and add the closing naming of the principle Untested proposal
The brief wants a whole project kept whole (a full design, a performance) Do not split it into exercises; put the clue inside the project brief and run question-led check-ins at its stages Untested proposal
The work goes to a real audience or client, or has non-negotiable constraints (safety, codes, accessibility) State the constraints directly at the start; explore within them; check the final work against them explicitly Untested proposal
The brief reports an error learners actually made (for example, symmetry everywhere read as balance) Put a work showing that error beside one that does not and ask what differs, before naming the principle Erroneous examples improve conceptual understanding by forcing comparison with correct models. [+M] (children, decimals; carried to design critique)
The brief asks for no test or quiz The check is the learner's own explanation of their work and of one new case during the critique; add nothing graded Untested proposal

Nine of the nineteen rows cite a claim. Only the domain-general, age and expertise rows rest on syntheses that test a comparison; the adult, online and error rows carry findings to learners or settings they did not test, and say so; the large-group row cites a claim for its limit; the rest are untested proposals.

Observation, state and explanation

What can be observed: the attempt (each version of the work), what the learner says about it when asked, which clue or term they use, what question the teacher asked and what the learner did next, and the learner's handling of a new case. "Has understood the principle" and "is copying the clue" are inferred states. A fluent oral explanation shows the learner can talk about this work; it does not show they can use the principle on a case the clue did not carry.

Response Plausible explanations Observation that could change the interpretation
Uses the clue's vocabulary fluently but cannot handle a new case The clue carried the idea; the learner learned words without the principle; the new case differs more than intended Give a new case close to the clue and one far from it; ask what to change and why, with sketching allowed.
Keeps producing near-copies of the clue Does not yet see what the clue is an example of; plays safe for the critique; time ran short Ask for a second version that keeps the principle but looks nothing like the clue; compare the two.
Stalls during making and waits for the teacher Too little prior knowledge to attempt; afraid of a public weak result; the problem is too open Offer a narrower version of the problem and a private first attempt; if they still cannot begin, show one example and its principle, then let them try.
Goes quiet or resists when the teacher speaks during group work The teacher interrupted a line of reasoning; the remark closed the discussion; the group was off task Note whether the group's talk resumes after the teacher leaves, and what the teacher said just before.
Explains the principle well in the closing, makes the same mistake next week Understanding without the habit of checking; interference from a new clue; the naming came too late to connect to the work Ask the learner to check last week's principle in this week's work before the critique.

These rows are proposals, untested with learners as diagnostics. Asking a question is itself guidance and changes what it observes; record what was asked, and keep the explanations that remain open.

Evidence and qualifications

No claim in the wiki compares clue-and-question guidance with direct explanation of the same principle on a later outcome. The claims below bear on parts of the relationship.

Learners, domains and outcomes differ across these claims (children learning a science strategy, adolescents in deliberation, eighth graders with tutors, meta-analyses across domains), so do not rank them by effect label. None is from a design studio except the last, and none measures delayed retention of understanding gained by exploring against understanding gained by being told.

The evidence by question

Further evidence, not yet read against this model

Objective and learner-valued goal

The designer's objective is usually that learners can see and use a principle in new work, explain why, and later work without the clue. Elicit separately what learners want: to produce a work they are proud of, to pass the jury, to be told what the teacher wants, to avoid being exposed in critique. A learner who values a good mark may ask for the "right answer" and read questions as withholding; one who values their own voice may resist the closing explanation. Agreement and divergence are both observations, and a request to "just tell us" is information about the learner's goal, not only about their understanding.

For example, a first-year architecture student may want to make a striking model for the end-of-term jury, while the studio's objective is that they can explain the principle of hierarchy in any composition. A question-led critique serves the objective; letting them name what made their model striking, and then naming the principle from it, serves both. Record both, so that a good jury result is not read as the understanding the course was aiming at.

What would revise this model?

The expectation should weaken if comparisons of question-led against explanation-led guidance during the same exploration, with comparable learners and an outcome on new cases, show no advantage, or show that the advantage disappears at a delay. If teachers prepared to ask questions still mostly interject explanations (as the recorded episodes suggest unprepared teachers do), the model becomes a teacher-preparation problem before it is a learner one. If naming the principle at the end adds nothing beyond critique, drop step 6; if naming it first and then critiquing matches the default, the order matters less than the critique. Do not protect the model by calling every failure "too much telling" after the fact.

Understanding shown in an oral explanation of one's own work, judgement of a new case, independent design work later in the course, and lasting understanding are separate claims. The present evidence establishes no amount of exploration time, no number of clues and no hand-over criterion.

Claims

Examples

Key Sources

  • Land, Susan M.; Hannafin, Michael J. (1996). Student-Centered Learning Environments: Foundations, Assumptions, and Implications. https://eric.ed.gov/?id=ED397810
  • Sözen, G., & Özen-Yavuz, A. (2024). Proposing a course schedule for architectural basic design studio guided by Bloom's Revised Taxonomy. Turkish Journal of Education, 13(4), 379-409. https://doi.org/10.19128/turje.1495965
  • Johnson, David Kenneth. (1995). CRITO: Informal Logic, Critical Thinking, and the Argumentative Essay. https://eric.ed.gov/?id=ED384062
Page metadata
Field Value
type principle
id understanding-cultivated-not-described
aliases instructor-indirect-narration-studio, impersonal-critique-provides-emotional-space
title Understanding must be cultivated through exploration, not described or told
description For a learner who can make attempts at a new concept or principle, exploration guided by clues, questions on the learner's own ideas and critique of everyone's work, with the principle then named from that work, is expected to deepen understanding more than being told first or exploring unguided, though no claim tests indirect against direct guidance.
status review
canonical True
generated by claude/unspecified on 2026-10-15
trust unverified — no one has confirmed this page's content
sources 3 — land-1996, sözen-2024, johnson-1995

Frontmatter is the page's machine-readable record — see the Schema & Guide.