Highlight Smarter: What to Highlight and Why

What to Highlight: The Four Categories

Most people highlight too much. They mark anything that seems interesting, ending up with pages of yellow that offer no more guidance than unmarked text. Research consistently shows that heavy highlighters remember no better than non-highlighters. The problem isn’t highlighting itself — it’s indiscriminate highlighting.

What to highlight while reading comes down to four categories:

Core arguments: The main claims the author is making. Not the setup, not the examples — the actual thesis and key supporting points. If you had to explain what this text argues in two sentences, what would you quote?

Surprising facts: Information that challenged your existing beliefs or taught you something you didn’t know. If nothing surprises you, either you already knew this material or you weren’t reading actively.

Key term definitions: Specialized vocabulary or familiar words used in specialized ways. These are the words you’d need to understand to discuss this topic with an expert.

Passages you’ll revisit: Quotes you might use in your own writing, reference points for future projects, or ideas you want to develop further. The test: will you actually come back to this?

Everything else — context, transitions, examples that illustrate without adding new information — can stay unmarked.

The Prompts: Before and After

PR019: Identify Before Marking

The Key Term Identifier helps you spot what’s actually central before you start marking. Ask AI to identify which words are central to understanding, which might appear in similar texts, and which have specialized meanings. This creates a mental filter: when you see these terms, they’re candidates for highlighting. When you don’t, they’re probably not.

The ranking by importance is particularly useful. Not all key terms are equally important — some are foundational concepts, others are supporting vocabulary. Highlight the foundational ones first.

PR035: Validate After Marking

The Highlight Validator checks your work. Share the original passage and your highlights, and ask AI to evaluate: Did you capture the core? Did you miss crucial evidence? Did you highlight too much setup?

This feedback loop trains your judgment. After a few rounds, you’ll naturally start making better selections without needing the validator.

Examples: Good vs Bad Highlighting

Bad highlighting: Marking entire paragraphs. Highlighting introductory phrases like “Research shows that…” or “It’s important to note that…” Marking anything that sounds impressive without checking if it’s actually the core claim.

Good highlighting: Marking just the specific claim within a longer paragraph. Highlighting the number or finding, not the framing around it. Marking the term being defined, not the full definition (you can reconstruct context from the term).

The test: if you looked at only your highlights, could you reconstruct the main argument? If yes, you highlighted well. If you’d need the surrounding text to make sense of them, you highlighted too narrowly. If the highlights alone feel overwhelming, you highlighted too much.

The Key Takeaways vs Key Quotes prompt (C019) can help distinguish between passages worth paraphrasing (takeaways) and passages worth preserving verbatim (quotes for highlighting).

Building the Habit

Selective highlighting is a skill that improves with practice. Start by deliberately under-highlighting — you can always add more on a second pass, but removing mental clutter is harder. Use the prompts to calibrate your judgment, and over time you’ll internalize what’s worth marking.

You’ve now completed the Notes & Memory pillar. For the complete toolkit, return to the pillar page or explore the full AI for Reading hub.