ASL-LEX: A Lexical Database for ASL
- Jul 13
- 3 min read
Updated: Jul 17
For spoken languages like English, researchers have long had detailed reference lists of words: how common each word is, how long, how it's built from sounds. These databases are behind a huge amount of what we know about how people read, learn, and recognize words. Sign languages didn't have anything like that at this scale. That gap held back research, and it held back technology too. You can't study patterns across a language, or build good tools for it, if you don't have an organized description of its words.
ASL-LEX filled that gap for American Sign Language. Naomi Caselli, Zed Sevcikova Sehyr, Ariel Cohen-Goldberg, and Karen Emmorey built it. It's free, it's public, and it has become a foundation for a lot of the sign language research and technology you'll read about elsewhere on this site.
From about 1,000 signs to 2,723
ASL-LEX came out in two big steps.
The first version, published in 2017, catalogued information about nearly 1,000 ASL signs. For each one, it recorded how common the sign is (frequency ratings from deaf signers), whether it looks like what it means (iconicity ratings), how long the sign and its video are, its grammatical class, and whether it's initialized, a fingerspelled loan sign, or a compound. It also coded the phonological building blocks of each sign, the parts a sign is made from, and used those to estimate how signs relate to one another. From the start, there was an interactive website, asl-lex.org, where anyone could browse and download the database. The second version, ASL-LEX 2.0, published in 2021, expanded the map to 2,723 signs and added more lexical and phonological detail for each one. It's the version most people use today.
What's in it
ASL-LEX 2.0 describes 2,723 ASL signs. For each one, it records a rich set of facts. A few of the important ones:
How the sign is built. Every sign gets a detailed phonological description: its handshape, where on or near the body it's made, its movement, and so on. Signs, like spoken words, are built from these smaller parts.
How common it is. Deaf ASL signers rated how often they see and use each sign. That "frequency" matters a lot, because common words behave differently from rare ones across every language.
How iconic it is. Some signs look like what they mean, and some don't. ASL-LEX includes ratings of that resemblance, from deaf signers and from people who don't sign. It even includes "guessability": can someone who doesn't know ASL guess what a sign means just by looking?
And more. Sign and video length, lexical class (like whether it's a noun or verb), whether a sign is fingerspelled or built from a letter, and measures of how phonologically simple or complex it is.
You don't have to be a researcher to use it. On asl-lex.org you can explore the whole lexicon visually, search it, and see how signs relate to one another.
Why this matters
ASL-LEX is a piece of infrastructure. For research, it lets scientists study how deaf children learn signs, how signers recognize and produce them, and how ASL is organized as a language, with the same rigor long available for spoken languages. Several other Deaf Center studies lean directly on it. For technology, ASL-LEX has become a backbone. When researchers teach a computer to recognize signs, they often use ASL-LEX's descriptions of how each sign is built. The sign datasets and models described in our other posts (like Sem Lex and ASL Citizen) connect back to this database. It's part of why sign language technology can treat ASL as a structured language rather than a set of unexplained gestures.
Key takeaways
ASL-LEX is a free, public, searchable database of American Sign Language signs.
It grew from about 1,000 signs in the 2017 release to 2,723 signs in ASL-LEX 2.0 (2021), with more detail added along the way.
For each sign it records how it's built (handshape, location, movement), how common it is, how iconic it is, and more.
Deaf signers' own ratings of frequency and iconicity are central to it.
Looking across the whole lexicon, frequent signs tended to be less iconic and simpler than rare signs.
ASL-LEX has become a foundation for sign language research and technology, including the sign-recognition tools described in our other posts. Explore it at asl-lex.org.
Sources: Caselli, N. K., Sehyr, Z. S., Cohen-Goldberg, A. M., & Emmorey, K. (2017). ASL-LEX: A lexical database of American Sign Language. Behavior Research Methods, 49(2), 784–801. https://doi.org/10.3758/s13428-016-0742-0. And: Sevcikova Sehyr, Z., Caselli, N., Cohen-Goldberg, A. M., & Emmorey, K. (2021). The ASL-LEX 2.0 Project: A Database of Lexical and Phonological Properties for 2,723 Signs in American Sign Language. The Journal of Deaf Studies and Deaf Education, 26(2), 263–277. This post summarizes both papers for a general audience, drafted from the published abstracts and introductions. Reviewed for non-deficit language.
