Data & indices
LINGACON maintains two headline measures. ALPS counts the languages in people's repertoires. The GLCI measures what those languages achieve: whether one person can understand another. Both are built country by country from documented sources, under one set of definitions, and published openly with the papers that describe them.
ALPS
Average Language a Person Speaks
ALPS is the population-weighted average number of languages spoken per person. "Speaking" is fixed at one explicit threshold, conversational competence at CEFR A2 and above: enough to handle an everyday exchange, more than a memorized phrase and less than native-level mastery. Fixing that threshold is what makes national figures comparable; sources that use looser or stricter criteria are harmonized to it, and the adjustment is documented.
Two design choices matter most. First, the index counts an unlimited number of languages per person. Many instruments stop at a first and second language, or cap the count at three or four; the Eurobarometer, for example, asks about a limited set of foreign languages, so anyone beyond that ceiling is truncated. A person who speaks six languages counts as six. Second, every national figure is checked against independent evidence instead of being taken from a single source at face value.
The current global estimate is 1.57 (band 1.42–1.70), across 237 countries and territories, with 934 languages indexed. ALPS is reported to two decimal places. The highest value backed by a direct survey is Luxembourg's, 3.33.
GLCI
Global Linguistic Connectivity Index
The GLCI asks a different question: take two people at random; what is the probability that the one listening understands the one speaking? Aggregated over all pairs, it measures how linguistically connected a population is. The global value is 11.3%. For two people from the same country it is about 63%; for two people from different countries it is 7.5%. English carries 39% of the global value.
The index does not treat languages as isolated boxes. Someone who speaks Spanish understands a good deal of Portuguese; a Czech speaker follows much of Slovak; speakers of neighbouring Arabic varieties understand each other to a degree that depends on which varieties. LINGACON accounts for this through a mutual intelligibility coefficient (MIC), a documented matrix of partial comprehension between related languages. Where intelligibility has been measured, the matrix uses the measured value; other pairs carry assessed values under written rules, marked as such. The matrix is asymmetric, because comprehension often is.
The methodology, the coefficient matrix with its sources, and the sensitivity of the headline figure to alternative conventions are published with the GLCI paper.
Coverage and sources
The dataset covers 237 countries and territories, sovereign states and dependent territories alike, home to 8.2 billion people (UN World Population Prospects 2024). 216 of them carry language-by-language detail. In the global GLCI the other 21 count at zero connectivity; the cross-country figure is computed over the 216.
National figures draw on population censuses, large survey programmes such as the Eurobarometer and the Afrobarometer, national statistical offices, and documented language surveys, harmonized to common definitions. No single source is taken at face value: instruments that appear to ask the same question often produce different answers because they apply different thresholds, and reconciling them is a large part of the work.
Open data and code
The data and code behind the ALPS and GLCI papers are deposited on the Open Science Framework: supplementary materials, row-level data, the pair table and the analysis code. Licences are stated for each component.
Detail beyond the deposit, such as sub-national breakdowns or named country groups, and uses the deposit licences do not cover can be discussed at research@lingacon.com; see also Partnerships. Corrections are welcome at the same address.