A Formula Can Survive Translation. A Problem Cannot Always.

Why mathematical symbols travel easily between languages while mathematical meaning must be reconstructed

A formula can cross a linguistic border almost unchanged. A problem must cross through a human mind.

— Tymur Levitin

There is something extraordinary about this expression:

x² + y² = z²

Write it in English.

Spanish.

German.

Ukrainian.

Polish.

The symbols remain recognizable.

The relationship survives.

Now try doing the same thing with a mathematical problem.

Something changes immediately.

The numbers may remain identical.

The mathematical relationship may remain identical.

The correct answer may remain identical.

But the learner no longer receives the mathematics directly.

They receive a linguistic representation of a mathematical situation.

And before they can solve the mathematics, they must reconstruct the situation.

Mathematics Has More Than One Language

We often say that mathematics is a universal language.

There is truth in that statement.

But it hides an important distinction.

Mathematics does not exist in only one representational system.

Consider the same relationship expressed in several ways:

25% of 48

0.25 × 48

A shaded quarter of a diagram.

A sentence:

Twenty-five percent of forty-eight is twelve.

Or in Spanish:

El veinticinco por ciento de cuarenta y ocho es doce.

These representations point toward the same mathematical relationship.

But they are not cognitively identical.

A learner may understand one immediately and struggle with another.

That difference tells us something important.

Understanding mathematics includes the ability to move between representations.

Symbols Compress Relationships

Mathematical notation is extraordinarily efficient.

Consider:

3x + 4 = 19

Very little linguistic interpretation is required for someone familiar with algebraic notation.

The symbols encode a relationship compactly.

Now expand it:

Three times a number plus four equals nineteen. Find the number.

The mathematical structure is still there.

But language has introduced additional processing.

Now Spanish:

El triple de un número más cuatro es igual a diecinueve. ¿Cuál es el número?

Again, the underlying mathematics has not changed.

But access to it has.

The learner must recognize that:

el triple de un número

corresponds to:

and that:

más cuatro

corresponds to:

+4.

The learner is translating—not necessarily between Spanish and English, but between representational systems.

This Is Not Ordinary Translation

That distinction matters.

Suppose a student sees:

la raíz cuadrada de 81

and knows that raíz cuadrada means square root.

That is lexical translation.

But consider:

La edad de Ana es el doble de la edad de Luis menos tres años.

Now the student must construct a mathematical relationship from language.

What does el doble de modify?

Where does menos tres años belong?

Which quantity should become the variable?

The problem is no longer:

What does this Spanish word mean?

It is:

What mathematical structure does this sentence describe?

That is a much more sophisticated cognitive operation.

A Word Problem Is a Model-Building Task

Before solving many word problems, a learner must build an internal model.

A situation is described.

Some information matters.

Some may not.

Quantities have relationships.

One quantity is unknown.

The student must decide what mathematical structure represents the situation.

Only then does calculation begin.

This means a word problem has at least two major stages:

Situation → Mathematical model → Solution

When another language is involved, another layer appears:

Linguistic input → Situation → Mathematical model → Solution

If the learner fails at the first transition, the mathematics may never get a chance to operate.

This Explains a Strange Classroom Phenomenon

A student solves:

0.30 × 80

without difficulty.

Then struggles with:

Una chaqueta cuesta 80 euros y tiene un descuento del 30 %. ¿Cuánto dinero se descuenta?

We might say:

The student doesn't understand percentages.

But perhaps they understand percentages perfectly.

Maybe they do not know descuento.

Maybe they interpret cuánto dinero se descuenta incorrectly.

Maybe they understand the vocabulary individually but fail to construct the relationship among price, percentage and discount.

The mathematical answer alone cannot diagnose the failure.

Representation Can Hide Competence

This is why the same mathematical knowledge can appear strong or weak depending on how a problem is represented.

Give the learner an equation.

They succeed.

Give them a diagram.

They hesitate.

Give them a graph.

They succeed again.

Describe the same relationship verbally.

They fail.

Has their mathematics changed four times?

Probably not.

What changed was the route through which the mathematical structure had to be recognized.

This distinction becomes especially visible in multilingual education.

Another Language Adds Processing, Not Necessarily Ignorance

A learner working in a second language may need to:

read the sentence;

identify unfamiliar vocabulary;

resolve grammar;

determine relationships between quantities;

hold those relationships in working memory;

construct the mathematical model;

choose an operation;

calculate;

then formulate an answer in the target language.

Compare that with seeing:

48 × 0.25

The difference in processing demand is obvious.

This is why slower performance in mathematics through another language should not automatically be interpreted as weaker mathematical ability.

Sometimes the learner is simply doing more work before reaching the mathematics.

But We Should Not Blame Language for Everything

The opposite mistake is equally dangerous.

A learner may understand every word in the problem and still fail to construct the mathematical model.

Then the difficulty is mathematical.

Perhaps they do not understand percentages.

Perhaps proportional reasoning is weak.

Perhaps they cannot distinguish additive from multiplicative relationships.

Language support alone will not solve that problem.

The important question is therefore not:

Is this a math problem or a language problem?

It may be either.

It may be both.

The better question is:

At which transition does understanding break?

The Transition Is Where Learning Becomes Visible

Consider this chain:

Words → Situation → Relationship → Representation → Operation → Answer

A learner can fail between any two stages.

They understand the words but misread the situation.

They understand the situation but cannot identify the mathematical relationship.

They recognize the relationship but cannot represent it algebraically.

They create the correct equation but cannot solve it.

They solve it but cannot explain the answer.

These are different educational problems.

A single mark such as incorrect hides all of them.

This Is Why Correct Answers Can Also Mislead Us

The reverse is possible.

A student may produce the correct answer without having built a robust mathematical model.

Perhaps they recognized a familiar exercise pattern.

Perhaps they applied the operation that usually appears on that worksheet.

Perhaps they guessed correctly from the available numbers.

This connects to another problem I explored in:

A Student Can Get the Right Answer and Still Not Understand the Mathematics

https://medium.com/@timurlevitin/a-student-can-get-the-right-answer-and-still-not-understand-the-mathematics-2fc7d9d19113

Correctness is important.

But mathematical understanding involves more than arriving at the expected number.

It includes knowing why that operation represents the situation.

A New Language Can Expose Hidden Mathematical Habits

This is one of the unexpected advantages of learning a familiar subject through another language.

The new language slows down processes that previously felt automatic.

A learner may suddenly notice:

Why does this phrase imply multiplication?

Why does that expression represent a ratio?

Why does at least change the inequality?

Why does el doble de create a multiplicative relationship?

Why does por cada express a rate?

The linguistic disruption can expose mathematical assumptions that had become invisible.

Something that initially makes learning harder can eventually make understanding more explicit.

Translation Can Therefore Become Analysis

At a beginner stage, translation may look like:

equation = ecuación

percentage = porcentaje

square root = raíz cuadrada

Useful.

But deeper bilingual mathematical work asks different questions:

How is the relationship encoded?

Which part of the sentence corresponds to which part of the equation?

Does another language package the relationship differently?

Which grammatical structure signals comparison?

Which expression signals proportionality?

Now translation is no longer vocabulary substitution.

It becomes structural analysis.

This Is Where Language and Mathematics Truly Meet

The interesting part of Language + Subject education is not simply teaching subject terminology in another language.

It is examining the point where linguistic structure carries disciplinary meaning.

In mathematics, this may be:

condition;

comparison;

quantity;

sequence;

ratio;

change;

probability;

logical consequence.

Expressions such as:

si... entonces...

mayor que

menor que

por cada

al menos

como máximo

are not decorative Spanish surrounding the mathematics.

They help construct the mathematical relationship.

The Formula Is Already a Translation

There is an even deeper point.

When we write:

d = vt

we have already translated.

A physical situation has been compressed into symbolic form.

When we draw a graph, we translate again.

When we describe the graph verbally, we translate again.

When we turn a verbal problem into an equation, we translate again.

Mathematical thinking constantly moves among representations.

So bilingual mathematics does not introduce translation into a process that previously contained none.

It makes an existing process more visible.

The Real Skill Is Representational Flexibility

A strong learner does not merely recognize one representation.

They can move.

Words → equation.

Equation → graph.

Graph → explanation.

Diagram → relationship.

Relationship → prediction.

One language → another language.

Each movement tests whether the underlying concept survives a change in form.

That is much closer to understanding than memorizing a single procedure attached to a familiar-looking exercise.

This Is Why Mathematics Can Be Powerful for Language Learning

The relationship also works in the opposite direction.

A learner who already understands mathematics possesses a rich conceptual framework.

New Spanish expressions can attach to existing mathematical meaning.

resolver una ecuación

is not learned in isolation.

The learner already knows what solving an equation means.

la raíz cuadrada

connects to an existing operation.

la probabilidad

connects to an existing concept.

This is the practical idea behind:

Learn Math in Spanish

https://timurlevitin.blogspot.com/p/learn-math-in-spanish.html

And it is why familiar academic knowledge can sometimes become an unusually efficient context for acquiring another language.

But Familiarity Does Not Remove the Need for Diagnosis

A learner may know the mathematics but lack Spanish.

Another may know Spanish but lack the mathematics.

Another may need both.

The same visible difficulty can emerge from completely different underlying structures.

Our first article in this cycle explored that problem directly:

The Numbers Didn't Change. So Why Did Math Suddenly Become Harder?

https://timurlevitin.blogspot.com/2026/08/the-numbers-didnt-change-so-why-did.html

The central principle is simple:

knowledge and access to knowledge are not identical.

Mathematics allows us to see the distinction with unusual clarity.

Perhaps Mathematics Is Universal in a Different Sense

Maybe mathematics is not universal because every human being experiences it without language.

We do not.

We encounter mathematics through symbols, diagrams, teachers, textbooks, examples, cultural conventions and words.

Perhaps its universality lies somewhere deeper.

Different representational systems can point toward the same underlying relationship.

Spanish can describe it.

English can describe it.

An equation can encode it.

A graph can display it.

A diagram can reveal it.

The representations differ.

The relationship can survive.

And learning becomes powerful when a person begins recognizing the relationship regardless of the form in which it arrives.