LINGACON is an independent research project whose overarching theme is the study of human cooperation. Its current work starts from two questions. How many languages does a person actually speak? And if you pick two people at random, what are the chances that one understands the other?
Both questions sound simple. We found no earlier global answer to either that could be compared from one country to the next, because the world's language data was collected for other purposes: surveys ask different questions, apply different proficiency thresholds and mean different things by "speaks". We set out a measurement standard that makes these sources comparable, applied it country by country, and published the results together with the data and code behind them.
The 11.3% counts every pair of people on Earth, neighbours included. For two people from different countries it falls to 7.5%; for two people from the same country it is about 63%. Put the other way round, in more than nine out of ten pairs of people from different countries, the one listening would not understand the one speaking. English alone carries 39% of the world's connectivity.
We think that many of humanity's hardest problems, from climate to conflict to pandemics, are at bottom problems of coordination. Knowledge and resources often exist; acting together at planetary scale is the hard part. Cooperation of any kind begins with the plain ability of two people to understand one another.
Language is one of the places where that ability is won or lost. Every additional language a person speaks is an advantage to that person and, quietly, a small addition to the connective tissue of humanity as a whole. Our long-term aim is to see that connectivity grow: better mutual understanding between people and, through it, better cooperation across domains and borders.
Measurement comes first, because what no one can see is hard to improve. So we build the numbers, publish them openly with the data behind them, and invite anyone to check them.
A note on where the science ends. Our indices measure how much connectivity there is. They are silent on how much there ought to be. That multilingualism is a good worth increasing is our premise, stated openly as a premise; it is not a finding derived from our data. We keep the two apart deliberately, and we would rather our measurements be used by people who disagree with us than bend them toward our own conclusions.
The LINGACON Multilingualism Assessment Standard: a five-level framework for measuring multilingualism comparably in a survey, a census or an organization, from one added question to a full national audit. It is open and aligned with the CEFR.
Average Language a Person Speaks: a population-weighted estimate for 237 countries and territories, built from censuses, survey programmes and national statistics, and harmonized to one proficiency threshold, conversational ability at CEFR A2 and above. 934 languages are indexed.
The Global Linguistic Connectivity Index looks at what those languages achieve. It is the probability that a randomly chosen listener understands a randomly chosen speaker, within countries, across borders and worldwide, with partial understanding between related languages taken into account.
The three preprints below are on SocArXiv under CC BY 4.0. The data and code behind the ALPS and GLCI papers are deposited on the Open Science Framework.
The population-weighted mean repertoire across 237 countries and territories, with 934 languages indexed: 1.57 languages per person (1.568), within a band of 1.42–1.70.
Hovhannisyan, A. (2026a). The ALPS Index: A global estimate of the average number of languages a person speaks. SocArXiv. https://doi.org/10.31235/osf.io/3mtsx_v1
The probability that a listener drawn at random from the world's 8.2 billion people understands a speaker drawn at random: 11.3%, and 7.5% when the two live in different countries. English carries 39% of this connectivity.
Hovhannisyan, A. (2026c). GLCI (Global Linguistic Connectivity Index): How connected is the world? SocArXiv. https://doi.org/10.31235/osf.io/g3vxj_v1
What counts as speaking a language, how to ask about it, and how existing surveys ask. Version 2, which corrects the reading of the Afrobarometer Round 9 and Round 10 items on other spoken languages, is awaiting moderation; version 1 remains available.
Hovhannisyan, A. (2026b). The LINGACON Multilingualism Assessment Standard (L-MAS): Definitions, thresholds, and a review of survey practice. SocArXiv. Preprint page (latest version): https://osf.io/preprints/socarxiv/n96g3. Version 1: https://doi.org/10.31235/osf.io/n96g3_v1
Supplementary materials, row-level data, the pair table and the analysis code for the ALPS and GLCI papers. Licences are stated for each component.
Statistical offices can add a standardized multilingualism question to surveys they already run, and organizations, cities and regions can ask for an assessment under L-MAS. Researchers are welcome to replicate, criticize or extend the work; the data and code are open for that purpose.
The project takes no external funding. Partnerships and contributions are what let the work continue and stay open.