The Archive of the Indigenous Languages of Latin America

What the Maps Reveal: Patterns in AILLA’s Language Documentation

The Archive of the Indigenous Languages of Latin America (AILLA) at the University of Texas at Austin holds one of the largest collections of Indigenous language materials. AILLA’s holdings span more than 500 languages across 263 archival collections.1 The AILLA Language Atlas visualizes two of the most extensively documented language families in AILLA: the Mayan languages of Mesoamerica and the Quechua languages of the Andes. Each map plots languages geographically and orders them chronologically by the start of dedicated documentation in AILLA’s holdings. The result is a pair of interactive StoryMapJS visualizations that surface spatial and temporal patterns invisible when browsing AILLA’s collections language by language.

The Mayan map features 19 languages with individual slides (those with 100 or more items), while the Quechua map features 9 languages (those with 5 or more items). Both maps include trilingual support in English, Spanish, and Portuguese, and each language slide links directly to its corresponding page on the AILLA website. The maps are not comprehensive catalogs of every language in AILLA; they are curated narratives that highlight patterns in how these two families have been documented over time and across geography.

Geographic Concentration

The Mayan map immediately reveals that AILLA’s documentation clusters tightly around two regions: the highlands of Guatemala and the state of Chiapas in southern Mexico. Of the 33 Mayan languages represented in AILLA, 21 are spoken in Guatemala and 14 in Mexico, with three languages (Chuj, Popti’, and Tektiteko) spanning both countries.2 Belize, where four Mayan languages are spoken, has comparatively little representation. AILLA’s holdings reflect the geographic priorities of the researchers who deposited their materials, with the archive’s Mayan collections drawing heavily from fieldwork conducted in these two regions.

The concentration of documentation around Guatemala and Chiapas also reflects the institutional networks that fed into AILLA. Many of the largest collections came from researchers affiliated with institutions in the United States and Guatemala who conducted fieldwork in these two regions. The result is a map that reveals as much about the history of Mayan linguistics as a discipline as it does about the languages themselves.

The Quechua map tells a different spatial story. Quechua languages span five countries along the Andes, from Colombia in the north to Argentina in the south. Yet Peru dominates the holdings, with 11 of 17 Quechua languages spoken there.3 Ecuador accounts for six, while Colombia, Argentina, and Bolivia each have just one. The Quechua map thus reveals both the geographic reach of the language family and the uneven distribution of documentation across the countries where Quechua is spoken.

This distribution reflects where AILLA’s depositors conducted their research, not the relative vitality of Quechua across the region. South Bolivian Quechua, for example, has over two million speakers but only limited representation in AILLA’s holdings. Like any archive, AILLA’s coverage is shaped by the networks and fieldwork priorities of the researchers who contributed materials.

Temporal Patterns

The chronological ordering of slides exposes a striking pattern in the Mayan map. The earliest materials for 30 of the 33 Mayan languages date to the 1960s or earlier.4 The 1960s marked a period of intensive linguistic fieldwork in Mesoamerica, with researchers like Terrence Kaufman, Nicholas Hopkins, and Lyle Campbell conducting foundational studies across multiple Mayan language communities. Because the atlas distinguishes between the earliest item in AILLA’s holdings and the start of dedicated documentation for a specific language, slides are ordered by when substantial language-specific fieldwork began rather than by the earliest item date.5

Quechua documentation follows a different trajectory. The earliest materials are more evenly distributed across decades, from 1959 (South Bolivian Quechua) through the 2010s.6 Rather than a concentrated burst of fieldwork in a single decade, Quechua documentation in AILLA grew through many independent research projects over time. This pattern suggests a more gradual, decentralized accumulation of materials driven by individual researchers with distinct regional focuses rather than a few prolific field researchers documenting many languages simultaneously.

One of the most revealing temporal patterns is the gap between when materials were created and when they were deposited at AILLA. Across the Mayan languages, the average gap between the earliest recording date and the earliest deposit date is roughly 42 years.7 Nearly all deposits occurred after AILLA’s founding in 2000, with the earliest uploads in the system dating to 2002. This means that much of what AILLA holds was recorded decades before the archive existed. Researchers carried recordings, field notes, and transcriptions for years, sometimes an entire career, before depositing them. AILLA did not simply accumulate new documentation as it was produced; it became a repository for decades of prior fieldwork that had been sitting in researchers’ personal collections. The archive’s founding created a home for materials that might otherwise have remained inaccessible or been lost entirely.

Access and Restriction

The maps also surface patterns in access. Approximately 25% of items in the Mayan family and 22% in the Quechua family are currently restricted or embargoed, meaning they are not publicly accessible through AILLA’s website.8 These are not small numbers. Roughly one in four items in both families exist behind access controls. The percentage varies dramatically by language. Tzotzil has the most items of any Mayan language in AILLA (714), but only 25% of those items are publicly accessible. Ixil, the second largest collection (655 items), is only 37% publicly accessible. By contrast, nearly all of Ch’orti’s 340 items are publicly accessible.9

These restrictions are not arbitrary. AILLA operates under agreements with depositors and Indigenous communities that govern who can access certain materials. Restrictions may reflect community preferences, ongoing research embargoes, or the sensitive nature of the content. The variation in access across languages reflects the diverse agreements between AILLA, its depositors, and the communities whose languages are documented.

The atlas handles restricted materials carefully. One language in each family meets the item threshold but is not featured because the majority of its holdings are restricted. Tojolab’al on the Mayan map and Tena Lowland Quichua on the Quechua map. These languages appear instead in the closing summary slide with a note about their restricted status. Languages that are featured but have significant restricted holdings include a note in their slide text indicating how many items are publicly accessible. The goal is transparency. Researchers should know what exists and what they can access before navigating to AILLA’s website.

What the Atlas Reveals About AILLA

Taken together, these maps tell a story not just about the languages they represent, but about AILLA as an institution. The geographic patterns reveal where AILLA’s depositors have focused their work. The temporal patterns reveal how long it took for the materials to find a permanent home. The access patterns reflect the agreements that shape how an Indigenous language archive operates.

AILLA holds items in 562 languages.10 The maps reveal gaps among those that do. Languages with fewer than 100 items did not receive their own slides on the Mayan map, and the threshold for Quechua was set at just 5 items to ensure broader representation. Several languages in both families have zero or only one item in AILLA’s holdings. These gaps reflect the scope of materials that have been deposited at AILLA, not the full extent of documentation that may exist for these languages.

The AILLA Language Atlas is intended as both a discovery tool and an invitation. By making AILLA’s holdings visible through a spatial and temporal lens, the maps allow researchers, educators, and community members to see at a glance what the archive contains and where the gaps are. AILLA’s collections are not static. New deposits continue to arrive, and as the archive grows, new languages will cross the thresholds for inclusion. The maps are designed to evolve alongside the archive itself.

The geographic concentration of Mayan documentation in Guatemala and Chiapas, the 42-year average deposit gap, and the variation in access restrictions all reflect the specific history of how AILLA’s collections were built. By making these patterns visible, the atlas helps researchers and community members understand what AILLA holds and where to find it.

Notes

  1. “Archive of the Indigenous Languages of Latin America,” University of Texas at Austin, accessed March 2026, https://ailla.utexas.org. See methodology documentation for data extraction details. ↩
  2. AILLA Language Atlas dataset, languages_dataset.csv. Language metadata from AILLA’s API; item counts and date ranges from AILLA’s pre-migration database (AILLA2). See methodology documentation for details. Country counts are based on the “countries” field in AILLA’s metadata; some languages span multiple countries. ↩
  3. Ibid. ↩
  4. Ibid. Of 33 Mayan languages with at least one item in AILLA, 30 have an earliest_item_year of 1969 or earlier (6 in 1959, 24 between 1960 and 1969). ↩
  5. “Dedicated documentation” refers to substantial fieldwork focused on a single language (5 or more items per year or the beginning of a consecutive multi-year documentation effort), as distinct from comparative or analytical works that include data on multiple languages incidentally. ↩
  6. AILLA Language Atlas dataset, languages_dataset.csv. South Bolivian Quechua has the earliest items (1959); Chachapoyas Quechua is the most recently documented, with earliest items from 2018. ↩
  7. Ibid. Calculated as the mean of (earliest_deposit_year – earliest_item_year) across all 33 Mayan languages with both values present. Median gap is 42 years. ↩
  8. Ibid. Mayan: 1,456 of 5,816 total items are restricted or embargoed (25.0%). Quechua: 276 of 1,268 total items (21.8%). ↩
  9. Ibid. Tzotzil: 714 total items, 180 publicly accessible (25%). Ixil: 655 total items, 244 publicly accessible (37%). Ch’orti’: 340 total items, 338 publicly accessible (99%). ↩
  10. Ibid. ↩