ELAN Tutorial 1: Set Up Tiers and Annotate Sign Language Data

Added:

Critiques of Manual Tier-Based Glossing in Sign Language Linguistics

While ELAN is the industry standard for manual sign language annotation, many linguists and researchers criticize traditional "glossing" (using written spoken-language words to represent signs) as reductionist, phonocentric, and inadequate for capturing the multi-dimensional, spatial, and simultaneous nature of sign languages. Glosses can impose a spoken-language structure on sign data, leading to translation bias and a loss of non-manual and phonetic details. To counter these limitations, alternative perspectives advocate for phonetically-grounded transcription systems (like HamNoSys or SignWriting) and modern computer vision/machine learning frameworks. These advanced tools utilize automated pose estimation and continuous spatial tracking to analyze sign languages directly from video data, bypassing the subjectivity, labor-intensiveness, and linguistic limitations of manual, text-based ELAN tier annotation.

Basic concepts of sign language linguistics, particularly the structure of American Sign Language (ASL) and English glossing conventions.
An understanding of what linguistic annotation is and why multi-tiered transcription is necessary for analyzing non-verbal communication.
Familiarity with the phonological parameters of sign language, such as handshape, location, movement, and non-manual features.
General computer proficiency, including file path management and basic knowledge of digital video formats.
Defining tier dependencies, parent-child relationships, and linguistic types with constraints in ELAN for structured database design.
Creating and implementing Controlled Vocabularies (CV) to ensure coding consistency across large corpora.
Exporting annotation data from ELAN to statistical software (such as R, Python, or Excel) for quantitative linguistic analysis.
Using ELAN's advanced search and multi-file query functions to analyze complex patterns across multiple recording sessions.
Measuring inter-rater reliability (IRR) using ELAN's built-in annotation comparison tools.
42.4K views195likes12:57@EI435LingASLOriginal Release: 2011-06-22

ELAN is free linguistic analysis software used for transcribing and annotating video data, particularly for sign language analysis. To begin, users create a new project by selecting File > New and locating their media file. Tiers are annotation layers that organize different types of linguistic information (such as gloss, English translation, participant, annotator, lexical categories, and non-manual markers), and users can add new tiers by clicking Tier > Add New Tier and naming each tier according to their analytical needs. Annotation involves selecting the appropriate tier, dragging the cursor to mark the start and end points of a sign or feature, typing the annotation in all caps following established glossing conventions, and pressing Command+Enter (Mac) or Control+Enter (PC) to confirm. Users can adjust playback speed, volume, and subtitle display order to facilitate accurate transcription.