SLP-AA: Tools for Sign Language Phonetic and Phonological Research

Roger Yu-Hsiang Lo, Kathleen Currie HallInterspeech 2019

TL;DR

This paper introduces SLP-AA, a free, open-source software tool designed to revolutionize sign language research by enabling detailed phonetic transcription and sophisticated phonological analysis. It addresses the need for robust computational tools to study the intricate structure of sign languages, making complex linguistic research more accessible and efficient.

What Problem Does This Paper Address?

Sign languages, like spoken languages, possess complex linguistic structures. However, studying their phonetic and phonological aspects computationally has been challenging due to the lack of specialized tools. Existing transcription systems often lack the necessary phonetic detail, and performing systematic phonological analysis on sign language corpora is time-consuming and difficult. This paper tackles this gap by presenting SLP-AA, a software suite that bridges the divide between detailed phonetic transcription and advanced phonological analysis, aiming to facilitate a deeper understanding of sign language structure.

Methods

SLP-AA is built upon the Sign Language Phonetic Annotation (SLPA) system, a highly detailed method for transcribing hand configurations, including precise joint flexion/extension and finger separation. The software itself is implemented in Python 3 with a PyQt 5 graphical user interface and is available as open-source. It offers two primary modes:

  • Annotator Mode: Allows users to create phonetically transcribed corpora of sign languages.
  • Analyzer Mode: Enables users to perform phonological searches and analyses on these corpora.

A key novel feature highlighted is the extended finger search function within the Analyzer mode. This function allows users to search for signs based on specific finger configurations (e.g., extended index finger) or the number of extended fingers, offering both positive and negative search capabilities across multiple hand configurations.

Datasets

Not discussed in the paper. The paper focuses on the software's functionality and potential applications rather than specific dataset training or evaluation.

Key Insights and Ideas

  • Bridging the Transcription-Analysis Gap: SLP-AA effectively connects detailed phonetic transcription (SLPA) with powerful analytical capabilities, a crucial step for computational sign linguistics.
  • Granularity of Phonetic Detail: The SLPA system, as implemented in SLP-AA, captures a level of phonetic detail (e.g., individual joint movements) that is essential for nuanced phonological research, moving beyond holistic descriptions.
  • Empowering Phonological Research: The Analyzer mode, particularly the extended finger search, provides researchers with tools to systematically investigate phonological patterns, such as handshape markedness, which were previously laborious to study manually.
  • Open-Source Accessibility: Making SLP-AA free and open-source democratizes access to advanced sign language research tools, fostering wider adoption and collaboration.

Key Results

The paper does not present quantitative results in terms of accuracy or performance metrics. Instead, it demonstrates the utility of SLP-AA through the description of its features and potential applications. The key "result" is the availability of a functional tool that significantly streamlines and enhances sign language phonetic and phonological research. The extended finger search function is presented as a concrete example of how the software can accelerate tasks like verifying generalizations about handshape markedness, which previously required manual counting and analysis.

Limitations and Future Work

The paper acknowledges that the Annotator mode's features were presented at a previous conference (INTERSPEECH 2017) and refers readers to that publication. The focus here is on the Analyzer mode. While the paper highlights the potential of the extended finger search, it doesn't delve into the full range of possible linguistic applications or the software's performance on large-scale corpora. Future work would likely involve expanding the analytical functions, integrating with other linguistic resources, and potentially developing more advanced machine learning capabilities for sign language analysis.

Summary updated

BibTeX
@inproceedings{lo2019slp,
  author = {Roger Yu-Hsiang Lo and Kathleen Currie Hall},
  title = {{SLP-AA: Tools for Sign Language Phonetic and Phonological Research}},
  booktitle = {Interspeech},
  year = {2019},
}
Figure