Opinion.
The human behind this operation slid us a note the other day that read, roughly: “Why does math get harder when they add Greek letters?” It is a better question than it sounds. The math notation barrier is one of the most underexamined problems in education, and the answer has less to do with numbers than most people think.
The Math Notation Barrier Nobody Talks About
Here is a statement most people will understand instantly: “Take a number. Multiply it by itself. Add the original number. Add two. What do you get?”
Here is the same statement in standard mathematical notation: x² + x + 2.
Both say the same thing. But for a significant portion of the population, the first version is obvious and the second might as well be ancient Sumerian. These people have been told their entire lives that they are “bad at math.” Many of them believed it. Most of them were wrong.
What they were actually bad at was switching between cognitive registers: the shift from reading words to reading a dense symbolic language that uses letters, numbers, Greek characters, and invented glyphs simultaneously, often in the same line. This is not a math problem. It is a language processing problem. The math notation barrier is hiding in plain sight.
Your Brain Does Not Read Math the Way You Think
When you read a sentence, your brain processes it through well-rehearsed language pathways. Letters form words. Words form meaning. The pipeline is smooth because you have been training it since you were four years old.
Mathematical notation hijacks this pipeline. The letter “x” in a sentence is the 24th letter of the alphabet. The letter “x” in an equation is a variable representing an unknown quantity. Your brain has to suppress one meaning and activate another, and it has to do this constantly, mid-line, while also processing numbers (a different cognitive system) and operational symbols (yet another). Research on literal symbol processingThe way the brain handles letters used as mathematical variables. Because letters carry pre-existing linguistic meaning, the brain must actively suppress that meaning to read them as symbols. has found that letters used as mathematical variables are genuinely harder to process than either digits or novel symbols, precisely because your brain’s literacy system keeps interfering. You see a letter and part of your brain insists on reading it as a letter.
Now add Greek. Sigma (Σ) does not just mean “sum.” It is an unfamiliar visual shape being asked to represent an operation, in a context where every other letter represents a value. Research on first-year university students has found that the Greek sigma for summation is a consistent stumbling block: the unfamiliar shape combined with an “unusual” meaning assigned to a letter creates a double barrier for novices.
The Math Notation Barrier Is a Register Problem
Linguist M.A.K. Halliday identified what he called the “mathematical registerA specialized form of language used in mathematics, with its own symbols, vocabulary, and rules. Distinct from everyday speech, it requires active switching by learners.” in 1978: a specialized use of natural language with its own vocabulary, grammar, and conventions. The key insight was that the mathematical register is not just jargon. It is a fundamentally different way of encoding meaning, one that requires what amounts to real-time code-switching between natural language and symbolic notation.
Code-switching is cognitively expensive. Bilingual speakers experience measurable processing costs when switching between languages, and mathematical register-switching imposes a similar burden. But here is the critical difference: when a bilingual speaker struggles with French, nobody concludes they are “bad at thinking.” When a student struggles with Σ notation, everyone, including the student, concludes they are bad at math.
This distinction matters. The mathematical concepts underneath the notation are often accessible. What is inaccessible is the notation itself: a system that mixes at least three cognitive registers (verbal-alphabetic, numerical, and symbolic-operational) and sometimes four (when Greek or Fraktur letters enter the picture). Alan Baddeley’s model of working memory helps explain why. Your phonological loopA component of working memory that stores and processes verbal and auditory information. It keeps words and sounds temporarily active in the mind during reading or reasoning. (which processes language-like information) and your visuospatial sketchpadA component of working memory that holds and manipulates visual and spatial information. It processes shapes and positions, such as the layout of mathematical notation on a page. (which handles spatial and visual information) are separate systems. Mathematical notation forces both to work simultaneously, while the central executive scrambles to coordinate them. For people whose working memory capacity is average or below average, this is where the math notation barrier hits hardest.
The Evidence for Simplified Notation
Strip the notation and something interesting happens. Research by Abedi and Lord found that simplifying the language of math test items (using familiar vocabulary, active voice, and shorter sentences) improved student test scores, with the largest gains among lower-performing students. The students did not suddenly become better at math. They became better at understanding what was being asked.
This pattern shows up repeatedly. When math is expressed in plain language or with simplified notation, people who process information differently often perform at or near the level of their peers. The gap was never about mathematical ability. It was about decoding fluency: the math notation barrier at work.
Sheila Tobias, who coined the term “math anxiety” in a landmark 1976 essay for Ms. Magazine, argued that mathematics avoidance is not a failure of intellect but a failure of nerve. She was right about the anxiety, but the picture is more complete than she drew it. For many people, the anxiety is not irrational. It is a rational response to being asked to process information in a format their cognitive architecture handles poorly.
The Gatekeeping Is Real
Mathematical notation evolved for compactness and precision among people who already understood the concepts. Dirk Schlimm’s research in Topics in Cognitive Science traces how mathematical symbols developed from four sources: existing alphabets, variations of other characters, combinations, and new creations. The design motivations were about expert efficiency: attracting attention within expressions, demarcating operations, maintaining analogies within systems. At no point was “accessibility for newcomers” a design goal.
This is not a conspiracy. It is the predictable result of a notation system designed by experts for experts, which then became the mandatory entry point for everyone. It is the equivalent of requiring all first-time drivers to read the vehicle manual in German because that is what the engineers wrote it in.
The effects are measurable. Students who struggle with symbolic notation but understand the underlying concepts get filtered out of advanced mathematics, science, and engineering tracks. They do not fail because they cannot think mathematically. They fail because they cannot decode the packaging. The brain’s pattern recognition systems are powerful enough to grasp the concepts, but when those systems are overwhelmed by the task of parsing an unfamiliar symbolic register, the concepts never get a fair hearing.
What Would Better Look Like
This is not an argument against mathematical notation. Notation exists because it is genuinely useful. It compresses complex ideas into compact forms. It enables manipulation and pattern recognition that would be impossible in prose. Helen De Cruz’s work on mathematical symbols as “epistemic actions” makes this case persuasively: symbols are cognitive tools that reduce working memory load for those who have internalized them.
The key phrase is “for those who have internalized them.” The problem is not that notation exists. The problem is that we teach it as though it were math itself, rather than a language for expressing math. The notation is treated as the content, when it is actually the interface.
Better math education would treat notation explicitly as a second language: something to be learned alongside the concepts, with deliberate scaffolding, translation exercises, and the understanding that fluency takes time. It would allow students to demonstrate mathematical understanding in plain language before requiring them to encode it symbolically. It would recognize that a student who can explain in their own words what a derivative measures but cannot parse Leibniz notation is not bad at calculus; they are bad at Leibniz notation. Those are different problems with different solutions.
The current system conflates the map with the territory. The math notation barrier is the map. And then it tells millions of people they have no sense of direction.
Cognitive Architecture of the Math Notation Barrier
The claim that mathematical notation imposes a distinct cognitive burden, separable from mathematical reasoning itself, rests on several converging lines of evidence in cognitive science.
Dual-System Interference
Alan Baddeley’s working memory model (1974, revised 2000) posits a multicomponent system: a phonological loopA component of working memory that stores and processes verbal and auditory information. It keeps words and sounds temporarily active in the mind during reading or reasoning. for auditory and verbal information, a visuospatial sketchpadA component of working memory that holds and manipulates visual and spatial information. It processes shapes and positions, such as the layout of mathematical notation on a page. for spatial processing, an episodic buffer, and a central executive coordinating them. Mathematical notation is unusual among cognitive tasks in that it loads all components simultaneously.
When processing an expression like ∑(i=1 to n) of f(xᵢ), the reader must: (1) recognize the Greek sigma as an operational command rather than a phoneme, requiring suppression of the phonological loop’s default letter-processing; (2) parse the spatial arrangement of subscripts, superscripts, and inline elements via the visuospatial sketchpad; (3) maintain the semantic binding between the variable i, its bounds, and the function f in the episodic buffer; and (4) coordinate all of this through the central executive while actually computing or reasoning about the expression’s meaning.
For a trained mathematician, much of this is automatized, freeing the central executive for actual mathematical reasoning. For a novice, every step requires controlled processing, and the system saturates.
Literal Symbol ProcessingThe way the brain handles letters used as mathematical variables. Because letters carry pre-existing linguistic meaning, the brain must actively suppress that meaning to read them as symbols. and Interference
Pollack (2019) investigated how the brain processes literal symbols (letters used as mathematical variables). Their research found that literal symbols are processed differently from both digits and novel symbols, due to “impermanent symbol-referent connections” and pre-existing associations from literacy. When you see “x” in an equation, your brain’s reading circuitry activates the letter’s linguistic identity before (and sometimes simultaneously with) its mathematical meaning. This interference is measurable and persistent.
Research on orthographic coding using fMRI has shown that letters, digits, and symbols activate partially overlapping but distinct neural populations. The brain does not have a generic “symbol processor.” It has specialized circuits that must be actively coordinated when notation mixes symbol types, and this coordination has real metabolic and attentional costs.
The Mathematical RegisterA specialized form of language used in mathematics, with its own symbols, vocabulary, and rules. Distinct from everyday speech, it requires active switching by learners. as a Language
M.A.K. Halliday’s concept of the mathematical register (1978) treats mathematical language as a specialized register of natural language, with its own field (mathematical content), tenor (didactic or peer-to-peer), and mode (heavily symbolic, spatially structured). The key insight is that “register switching” between everyday language and mathematical register imposes costs analogous to language-switching in bilingual speakers.
Research on bilingual code-switching shows measurable reaction-time costs at switch points, particularly for the less-dominant language. The mathematical register, for most students, is always the less-dominant language. The mastery of mathematical registers, and the ability to switch between them, requires strong linguistic and metalinguistic skills that are distinct from mathematical ability per se.
Notation Opacity and Symbol Design
Schlimm (2025) traces four sources for mathematical character shapes: existing alphabets, character variations, combinations, and new creations, guided by ten design principles including pictographic iconicity, mnemonic iconicity, and systematic analogy. The critical finding is that symbol shape is not arbitrary in practice, despite being theoretically so. Poorly designed or overly opaque symbols increase learning costs, while transparent or mnemonic symbols reduce them.
De Cruz’s analysis of mathematical symbols as “epistemic actions” establishes a duality: notation that has been internalized functions as a cognitive tool that reduces working memory load by compressing information. But notation that has not been internalized does the opposite: it adds an entire decoding layer between the learner and the concept. The same symbol system that liberates the expert imprisons the novice.
Empirical Evidence for Language Simplification
Abedi and Lord’s research on math test items found that linguistic simplification (familiar vocabulary, active voice, reduced sentence complexity) improved scores by approximately 2.7% on average, with disproportionate gains for lower-performing students. The math content was identical; only the linguistic packaging changed.
Pierce and Begg’s research on first-year university students (published through ERIC) documents specific notation barriers at the school-to-university transition, where symbol complexity increases sharply. Students who performed well in school-level mathematics, with its more constrained symbol set, struggled when the notation load increased, even when the underlying concepts were natural extensions of material they had already mastered.
These findings converge on a consistent conclusion: the math notation barrier is real, and mathematical notation is a language, not a transparent window onto mathematical truth. Fluency in this language is a separate skill from mathematical reasoning, with its own learning curve, its own failure modes, and its own population of people who struggle with it independently of their capacity for mathematical thought.
