A17

Split-attention repair

Stop bouncing between a diagram and a separate wall of text, and get one integrated explanation that keeps eyes and mind on the same object.

When to use this: First time through the material
Learning impact: Clear help, modest gain
Learning evidence: Strong research tradition
Where you are with the topic: Meeting the idea
How often to use it: Useful once for a topic

When to use it

You are meeting this idea for the first time and need a way in.
The reading feels dense and you want a first map before the details.
The idea only makes sense when you can see it, and your notes are all prose.

What to ask AI

Type something like: “I have a diagram and a separate explanation, and they're splitting my attention.”.

Ask it to generate material for you to attempt, problems, cues, or a schedule, not finished answers.

Keep the hard part on your side: you retrieve, attempt, explain, or decide before you ask for help.

How it helps

You've seen the textbook page: figure on the left, explanation on the right (or worse, two pages later).

You read a sentence, hunt the matching label, lose the sentence, reread, hunt again.

That thrash is not a moral failing.

It is split attention, and cognitive load research treats it as a design defect, extraneous load that consumes the same working memory you needed for the actual idea.

AI cannot redraw your physical textbook, but it can produce an integrated verbal walkthrough of this figure: each element described in viewing order, labels bound to meanings in one stream.

The repair is temporary integration plus your own reconstruction.

Rank L in the register for a reason: narrow, and partially obsolete if you have true multimodal diagram tools (family M).

It still earns an article because most study material is still split, and the verbal integration trick remains useful on phones and in a11y contexts.

How to run it

Identify the split. Diagram + distant caption, slide + speaker notes, code + documentation in another tab.
Ask for an integrated narration in the order eyes should move.
Next: follow on the figure with a finger or cursor, do not read the narration alone.
Next: redraw or re-label the figure from memory.
Next: only then read any remaining prose for what the narration couldn't carry (caveats, proofs, exceptions).

Example exchange

You

I have a Krebs-cycle diagram and a separate paragraph, and I keep bouncing between them. Give me one walkthrough in viewing order, then stop so I can redraw from memory.

Start at the top with acetyl-CoA entering. The next label clockwise is citrate, that's the six-carbon compound formed first. Continue around until you hit where carbon dioxide leaves. When you're ready, hide both and redraw.

You

I redrew it. I blanked on where NADH is produced.

Mark that gap on your redraw. Look back only at that sector, then hide it and add NADH from memory.

Copyable prompt

I have a diagram and a separate explanation, and they're splitting my attention.
Give me ONE integrated walkthrough: describe the figure in viewing order, binding
each label to its meaning as we go. No separate "text section." Then stop and
ask me to redraw from memory.
DIAGRAM DESCRIPTION / LABELS: [ ]
SEPARATE TEXT: [ ]

The Tell

Here is how you know this method has flipped: you understand the paragraph and still couldn't teach the figure's structure with the text covered.

In that moment the AI (or the schedule, or the story) did the thinking, and you only recognised a finished product.

Tighten the prompt for split-attention repair so you attempt, decide, or retrieve before anything is handed to you.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated strong. The split-attention effect is a core cognitive load finding (Chandler & Sweller and subsequent CLT work): physically integrating diagram and explanation reduces extraneous load versus separate sources. Principle strong for the design fix; AI “integration” of text+diagram remains delivery-speculative and modality-limited in chat.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

The learning principle may be strong, but that does not prove an AI session delivers it well. Treat AI delivery as speculative unless a study of tutoring with this method is named, and none is claimed here.