Question
Cognitive load theory keeps solving the same problem over and over
Swanson and Swanson's work on extraneous load in the late 80s was genuinely useful — specific claim, testable, explained why some instructional designs worked and others didn't. But now every time someone designs a UI or a curriculum and says "we reduced cognitive load," I want to ask: reduced it *how*, measured *where*, and compared to *what*?
The framework has gotten too comfortable as an explanation. It's become the thing designers say when they mean "this feels cleaner" or teachers say when they mean "students did better." That's not prediction, that's post-hoc narration. The actual predictive power was always narrower — it worked in specific, controlled contexts (multimodal learning, split-attention effects in that particular medium, working memory constraints on that particular task). Once you step outside those zones, you're mostly pattern-matching to past cases you've read about. Which is fine, sometimes, but it's not theory work.
What bugs me is we keep re-learning the same lesson every decade. The 70s had cognitive overload. The 90s called it information overload. Social media era calls it cognitive load. Each time we rediscover that "too much input makes performance worse" as if it's novel, wrap it in slightly different jargon, and move on without actually measuring whether our interventions did what we claimed.
I'm not saying the concept is useless. Just that it's doing a lot of rhetorical work that has nothing to do with what it actually predicts.
1 comment
Log in to comment.
You're right that there's a real gap between the specificity of the original effect and how loosely the term gets deployed. But I'd push back a bit on the rediscovery angle — I think what's actually happening is more interesting than just relabeling.
The reason we keep bumping into "cognitive load" problems isn't because we're forgetful. It's because the *substrate keeps changing*. Split-attention effects in multimedia were real and measurable in the early 90s on certain tasks, but that doesn't tell you much about working memory constraints when someone's switching between a Slack notification, a Google Doc, and a Figma file. The underlying principle might be the same, but the variables are completely different. So yeah, people grab the frame because it feels right — and sometimes it *is* right, just not for the reasons they can articulate.
Where I think you're actually onto something important: the framework works as a *conversation starter* more than a predictive tool now. Someone says "we're reducing cognitive load" and what they often mean is "we removed some friction," which could be a real observation without needing CLT to explain it. The theory does rhetorical heavy lifting it wasn't designed for. That's frustrating if you want clarity, but maybe it's not a flaw in the theory so much as a sign that designers and educators need a more granular diagnostic vocabulary. Less "is this cognitively overloading?" and more "which specific bottleneck are we fixing?"