We know there are about 7,000 living languages on Earth. We know which ones are growing, which are dying, which have dictionaries and which have never been written down. Linguists have spent a century cataloguing every language.
But here is a question that nobody can answer: how many languages does the average person speak? Not just you an me. Not the average European. Not the average educated professional. The average human being, across all 8.3 billion of us.
Go ahead, try to find the number. Google it. Ask a linguist. Check the UN. Check UNESCO. Check the World Bank’s data portal. You won’t find it. Because nobody has ever measured it.
The number that doesn’t exist
Think about what we measure routinely. Global literacy rate? 87%. Global internet penetration? 67%. Average life expectancy? 73.5 years. These numbers exist because someone decided they matter, built a standard way to measure them, and then countries started reporting.
Multilingualism has no such number, no standard question, no agreed definition of “speaking” a language, no global framework. Some surveys ask “do you speak a second language?” Others ask “which languages do you use at home?” Others ask “rate your proficiency on a scale of 1 to 5.” The results are incompatible. A person counted as bilingual in a European survey might not qualify in an Asian census. You cannot compare the results. You cannot add them up. You cannot draw a global picture.
So we built one
The number is called ALPS©. Average Languages a Person Speaks. One number per country, per city, per organization. Comparable across borders and across time.
We estimate the current global ALPS at approximately 1.54. That means the average human speaks about one and a half languages. Most of the world is monolingual, or barely bilingual.
Some of you are already objecting. “In my country, everyone speaks at least three languages!” Maybe. Luxembourg’s ALPS is probably above 3.80. Singapore is 3+, and the Netherlands too. But China, Japan, Brazil, Russia, and the United States together hold over 3 billion people. Most of them speak one language. They pull the average down, hard. And that is exactly the kind of thing you can see only when you have a standard way to measure.
The Japan-Rwanda puzzle
Here is something that surprised us. Suppose you want to measure multilingualism in Japan (population 125 million) and Rwanda (population 14 million). Which country is harder and more expensive to survey?
The intuitive answer is Japan. Nine times the population. Must be nine times the work.
The real answer is Rwanda, with extreme linguistic diversity. Kinyarwanda, French, English, Swahili, and dozens of local languages in different combinations. A representative sample must be stratified across linguistic regions, urban and rural areas, education levels. The sample size might need 8,000 to 12,000 respondents.
Japan? Nearly everyone speaks Japanese. A standard sample of 2,000 to 3,000 tells you everything. Japan is linguistically simple. Rwanda is linguistically complex. Population is irrelevant. Diversity is everything. So measuring is three times harder. We did not expect this. But once we saw it, we could not unsee it.
What ALPS makes visible
When you have a single number that works the same way everywhere, patterns appear. You can compare countries. You can track a country over decades. You can measure whether language education policies actually work. You can tell a multinational company: “your São Paulo office has an ALPS of 1.4, and your Dubai office has an ALPS of 3.5, and that is why your cross-office projects keep failing.”
That last example is not theoretical. We built a sample assessment for a certain company, and the connectivity gaps between offices were dramatic. The bottleneck was not technology or management. It was language.
What happens now
We published the framework. It is called L-MAS, the LINGACON Multilingualism Assessment Standard. Five levels, from a single question (cost: nearly zero) to a full national audit. Every level produces the same core metric: ALPS. The paper is going up on PsyArXiv within days. The framework is free. The data will be free. The ALPS values for every country we can measure will be public and open. Because this number belongs to everyone, and it should have existed decades ago.
If you want to know more: l-mas.com
And if you are a government, a city, or an organization that wants to measure its own multilingualism — we do that too.
Areg Hovhannisyan is the creator of the Human Elements System (HES) and the LINGACON research program on global linguistic connectivity. He writes about measurement, human psychology, and the things we forgot to count.