Asking How and Why: Elaboration
Elaborative interrogation and self-explanation deepen encoding by multiplying the routes back to an idea — a fact woven into what you already know is reachable from many directions; an isolated fact from only one. · 10 min
The last unit was about arranging practice in time — when to study, in what order, how often. This one is about what happens in the moment you meet an idea: how to encode it so richly that it holds. The cheapest deep technique is also the oldest piece of advice, and most people skip it because it feels slow. When you learn a new fact, do not just take it in. Ask how it works and why it is true, until it connects to things you already know. That connecting is not decoration on top of memory. It is what makes the memory findable later.
Guess before you learn
You need to remember that the Netherlands reclaimed much of its land from the sea. Two ways to study it: (A) reread the sentence four times until it feels familiar; (B) ask and answer 'why would a country build land out of the sea?' and 'how would that even work?', connecting it to what you know about flooding and farmland. A week later, which sticks better?
The questions win. Tying a fact to what you already know — by asking how and why — stores it with several routes back, while rereading leaves it isolated and fading. If you chose rereading, you chose the technique that feels productive and ranks near the bottom of the evidence (folio 4).
9–12
3–5
A fact you only memorize sits by itself, and by itself it is easy to forget. But if you ask how and why until the fact links up with things you already know, it stops being alone. Suppose you learn that camels store fat in their humps. Ask why fat and not water — and now the fact connects to what you know about deserts and energy. Connected facts are far easier to remember.
Explaining to yourself works the same way. When you read a solved example, do not just nod along — say out loud why each step is done. The saying is what makes it stick.
6–8
Two techniques deepen how you store an idea. Elaborative interrogation means asking how and why about a new fact until it connects to what you already know: not just 'veins carry blood back to the heart' but why — and how that differs from arteries. Self-explanation means saying, in your own words, why each step of a worked example is correct, instead of reading it silently.
Both work by the same mechanism. A fact connected to your existing knowledge can be reached by many routes; an isolated fact has only one. Elaboration builds the extra routes, so when one path to the memory fails, another still leads back to it.
9–12
The evidence places both techniques at moderate utility. Dunlosky and colleagues' 2013 review — the same review that ranked rereading and highlighting near the bottom (folio 4) — rated elaborative interrogation and self-explanation as reliably helpful across many studies, though more demanding to use well. Chi and colleagues (1989) found that students who spontaneously self-explained worked examples learned far more than those who did not, from identical material.
Elaboration has a precondition: prior knowledge. Asking why helps most when you know enough to build a real answer; on wholly unfamiliar material the questions have nothing to connect to, and the benefit shrinks. This is why elaboration pairs naturally with the knowledge you already hold — and why understanding a subject tends to accelerate as it grows.
K–2
When you learn something new, ask why. Why does ice float? Because it is lighter than water. Asking why ties the new fact to what you already know, so it holds on.
One fact, all alone, is easy to lose. The same fact tied to three things you know is easy to find. Ask how and why, and tie it on.
Undergrad
The mechanism is integration. Elaborative interrogation (Pressley et al., 1987) and self-explanation (Chi et al., 1994) both prompt the learner to generate connections between new information and prior knowledge, embedding the target in a richer associative structure. On the elaborative-retrieval account met in folio 5, each added connection is another retrieval route: recall succeeds if any path reaches the target, so multiplying paths multiplies the chance of later access.
The precondition is domain knowledge, which makes elaboration self-accelerating. A learner with a developed schema generates more and better connections, gains more from each new fact, and thereby enriches the schema further. Novelty is the limiting reagent: elaborative interrogation of a fact you cannot yet situate produces shallow or mistaken explanations, and the gain collapses toward that of plain study.
Postgrad
Elaboration and self-explanation are generative encoding manipulations: the learner produces connective inferences rather than receiving them. Dunlosky et al. (2013) rate both as moderate-utility, constrained mainly by prior knowledge and by the quality of the explanations generated. Self-explanation's efficacy (Chi et al., 1989, 1994; meta-analysis by Bisra et al., 2018, g near 0.55) is largest when prompts push justification of underlying principles rather than mere paraphrase.
The account unifies with retrieval-based views: elaboration increases the number and diversity of cues from which the trace is reachable, raising the ceiling on retrieval strength by enriching storage. The failure modes are informative — illusory understanding from fluent but content-free explanation, and negative transfer from confidently wrong self-explanations — which is why generated explanations must be checked against a standard, not merely produced.
elaborative interrogation
Asking how and why about a new fact until it connects to what you already know, generating an explanation rather than just rereading the fact.
Why is this true?
Why does a fact connected to your prior knowledge survive better than an isolated one?
Because recall succeeds if any route reaches the memory. An isolated fact has a single path, so one failed cue loses it; an elaborated fact sits at the end of many paths, and any one of them can bring it back.
The shape says two useful things. First, the first few connections matter most — going from an isolated fact to one with three or four genuine links to prior knowledge does the bulk of the work. Second, the connections have to be real. An answer you make up that happens to be wrong, or a link too vague to mean anything, adds no route. The move is specific: take the new fact, ask why it is true and how it works, answer from what you actually know, and check the answer. A fact with several honest explanations attached is reachable from all of them.
Self-explanation is elaboration aimed at a worked example. Instead of reading a solved problem and nodding, you stop at each line and say why that step is allowed. It is more effort than following along, and Chi's studies found that the students who did it — often the stronger ones, but the habit can be taught — learned far more from the same page. Try it on the solution below: for each step, produce the reason before you read on. The reasons are not obvious just because the arithmetic is; naming them is what turns a watched solution into a method you own.
Self-explain each step: solve 2x + 3 = 11 — the steps fade as you master them
Subtracting 3 isolates the 2x term, and doing it to both sides keeps the equation balanced.
Dividing by 2 undoes the multiplication, leaving x by itself.
2(4) + 3 = 11 confirms the value satisfies the original equation.
Elaboration builds routes to an idea out of words. The next folio adds a second kind of route entirely — a picture. Pairing a clear diagram with the words lays down two memory traces where prose alone leaves one, and it turns out to help everyone, not only the people who believe they are visual learners.
Practice — new ink and old, interleaved
1.Which group predicted they would remember more?
2.From memory: give one reason retrieval strengthens memory more than rereading does.
Retrieval is effortful reconstruction — searching memory activates related knowledge and updates the memory's context, so the fact becomes reachable from more future situations.
How close were you? Grade yourself honestly — it sets your review date.
3.From memory: what is interleaved practice, and what happened in the 54-classroom trial?
Interleaving mixes related problem types in one session so each problem must be classified before it is solved. In Rohrer's 2020 randomized trial, interleaved classes scored 61% to the blocked classes' 38% one month later — d ≈ 0.83.
How close were you? Grade yourself honestly — it sets your review date.
4.From folio 5: after a week, Roediger and Karpicke's rereaders kept about 40% of the passage's ideas. About what percent did the repeated recallers keep?
5.From folio 2: how does connecting a new fact to an existing schema ease the narrow gate of working memory?
6.From folio 9: elaboration and interleaving are both effortful. What does interleaving practice that elaboration does not?
7.You need to carry a sixteen-character wifi password across the house in your head. In one sentence, what should you do to it first?
8.A worked example divides both sides of an equation by 4. In one sentence, self-explain why that step is allowed.
9.Feedback in hand, which wrong answer is most likely to be corrected on a retest?
10.From memory: state the 'many routes' explanation of why elaboration works, and one thing that can make an elaboration useless.
Elaboration ties a fact to prior knowledge, adding routes so recall can succeed from any of them. It fails when the connection is wrong or too vague to mean anything, or when there is no prior knowledge to connect to.
How close were you? Grade yourself honestly — it sets your review date.
11.You just read that warm air holds more water vapor than cold air. Which response is elaborative interrogation?
12.From folio 6: why does generating a wrong 'why' answer, then correcting it, often beat reading the right explanation cold?