TL;DR

Your hypothesis is partially true, but the framing matters:

It is not true that mixing a serif and a sans-serif inherently makes long-form text more readable. Modern typography research finds that the serif/sans distinction itself has little measurable effect on reading speed or comprehension on today's screens.

It is true, and broadly accepted by typographers, that giving headings and body text two identifiably different fonts improves navigation, hierarchy and visual rhythm — which is a different (and arguably more important) kind of "readability" for displays where people scan before they read.

The serif-for-body / sans-for-headings convention is one well-tested way to achieve that contrast, especially in books and long-form articles, but it's a convention, not a law. Two serifs, two sans, or a superfamily can work just as well.

So the practical principle isn't "always mix serif + sans." It's: use enough contrast between heading and body type to make hierarchy obvious, while keeping enough harmony that the page feels like one document.

1. What the typographers say

Robert Bringhurst — The Elements of Typographic Style

Often called the "Typographer's Bible". Bringhurst doesn't prescribe a serif/sans rule. His framing is about rhythm, proportion, and harmony, and about choosing typefaces that suit the content and reader. His typeface-combination advice is essentially: contrast must serve the text, not decorate it. The book itself is set in Minion (serif) for body and FF Scala Sans (sans) for secondary material — a serif-body / sans-headings pairing applied with discipline rather than as a default.

Matthew Butterick — Practical Typography

Butterick is one of the most-cited modern voices and explicitly contradicts the popular belief:

"You can mix any two fonts that are identifiably different. If you've heard that you can only mix a serif font with a sans serif font, it's not true."

He points out that American newspapers routinely set headlines and body in two different serifs, and argues that lower contrast between fonts is often more effective than higher contrast. His real rule is consistency of role: each font has a defined job (body, headings, captions, etc.).

Smashing Magazine — Best Practices of Combining Typefaces

Argues for the "tried-and-true combination of using a classic neutral heading typeface and a classic neutral body typeface," and notes you can get a strong, varied page from just two typefaces (e.g. Helvetica Neue + Garamond) by using different weights and sizes. Their core point: avoid letting two strongly-personalitied fonts compete; pick one to lead.

Book design tradition (e.g. Book Design Made Simple)

For long-form books, the dominant convention is serif body + sans-serif headings, with the body font usually leading. Reasons cited:

Serifs in body text create a horizontal reading rhythm that helps the eye flow across long lines.

Sans-serif headings are visually distinct at large sizes and don't clutter the page with serifs at display scale.

Matching x-heights between the two faces is more important than the serif/sans choice itself.

Recent trend — same-family serif + sans pairings

A more contemporary technique (e.g. Suisse Serif + Suisse Sans, Source Serif + Source Sans, IBM Plex Serif + Sans) uses serif and sans cuts of the same superfamily. This gives you the hierarchy contrast without any visual mismatch — a pragmatic middle ground that's become common in product UIs and brand systems.

2. What the reading-science research says

This is where the popular "serif vs. sans" debate falls apart. The empirical literature is genuinely mixed:

Older "low-resolution screen" finding: on early monitors, sans-serif was easier to read at small sizes because hinting/anti-aliasing couldn't render serifs cleanly. This is the origin of the "use sans on screen, serif in print" rule of thumb.

Modern high-resolution screens: this advantage has largely disappeared. Multiple studies (Bernard & Mills 2000; Boyarski et al. 1998; Josephson 2008) find no statistically significant difference in reading speed or comprehension between serif and sans-serif body text on screen.

Accessibility / low vision: a review of 18 studies in the Journal of Visual Impairment and Blindness concluded sans-serif (Arial, Helvetica, Verdana) is more readable for low-vision readers — a real exception worth noting.

Preference vs. performance: people often prefer sans-serif on screens but don't actually read faster or comprehend better with it. Preference and performance aren't the same metric.

The most honest summary, from researchers who've reviewed the field: the serif/sans choice has a much smaller effect on readability than font size, line length, line height, contrast, and the reader's familiarity with the typeface.

3. So why does pairing serif + sans for headings vs. body still feel right?

Because it solves a different problem than raw legibility:

Hierarchy at a glance. Two distinct font classes signal "this is a heading, this is body" before the reader processes any words. Size and weight do this too, but adding a class change makes it unambiguous.

Scanning behaviour. On screens especially, readers skim before they commit. Different families help headings act as visual landmarks.

Tonal contrast. Serifs read as more traditional/literary; sans as more modern/utilitarian. Pairing them lets the heading and body play complementary tonal roles.

Display vs. text optimization. Many serifs are designed to shine at body sizes; many sans are designed for display clarity at large sizes. Using each at its sweet spot is just good craft.

None of this requires the pairing to be specifically serif + sans. Two serifs (one display, one text) or a serif/sans superfamily work for the same reasons.

4. Practical principles for your work

For broad use across long-form reading and UI:

Default to two identifiably different fonts for heading vs. body when you want clear hierarchy with minimal effort.

Match x-heights between the two — this matters more than the serif/sans distinction. Mismatched x-heights make the page feel disjointed even if the fonts are "supposed" to pair.

Pick one font to lead (usually body, since there's more of it). The heading font supports the body, not the other way around.

Contrast through a combination of family, weight, and size — not just family. A bold sans heading over a regular serif body has three layers of contrast working together.

Consider a superfamily (Source Serif + Source Sans, IBM Plex Serif + Sans, Inter + a serif companion) when you want pairing safety with minimal risk.

Don't expect the pairing to make body text more readable. Choose the body font for legibility at your target size and line length first; choose the heading font to complement it second.

For screen-first products, body legibility comes from size, line-height, line-length and contrast — not from whether the font has serifs.

5. Key sources

Robert Bringhurst — The Elements of Typographic Style (4th ed., 2012)

Matthew Butterick — Practical Typography, "Mixing fonts" — https://practicaltypography.com/mixing-fonts.html

Jim Felici — The Complete Manual of Typography (general reference on pairing principles)

Smashing Magazine — "Best Practices of Combining Typefaces" (2010) — https://www.smashingmagazine.com/2010/11/best-practices-of-combining-typefaces/

Fiona Raeside-Elliott — Book Design Made Simple, "Combining serif and sans serif fonts" — https://www.bookdesignmadesimple.com/combining-serif-and-sans-serif-fonts/

Sarah Hyndman — Type Tasting, on serif vs sans serif font psychology and BBC Reith research — https://www.typetasting.com

Richardson, J.T.E. — The Legibility of Serif and Sans Serif Typefaces (Springer, 2022) — meta-review of the empirical literature

Bernard, M., Mills, M. et al. — multiple readability studies (cited widely in HCI literature)

Journal of Visual Impairment and Blindness — review of 18 studies on typeface legibility for low-vision readers

Rello, L. & Baeza-Yates, R. — readability research, including dyslexia-focused studies

The Readability Group / Bruno Maag — ongoing research on readable type, including the BBC Reith project

6. Where you might have read this originally

Given your background and what you describe, the most likely sources are:

Bringhurst's Elements — the canonical reference, most designers encounter the pairing idea here first

Butterick's Practical Typography — free online, widely shared in dev/design circles

Smashing Magazine's typography articles — common reading for web devs

Thinking with Type by Ellen Lupton — another common gateway, very accessible

If you remember an "important typographer" specifically and a longer-form treatment, Bringhurst is the strongest bet. If it was a punchy, web-friendly piece, Butterick.