Emphatic segments and emphasis spread in Lebanese Arabic: a Real-time Magnetic Resonance Imaging Study

Assaf Israel, Michael I. Proctor, Louis Goldstein, Khalil Iskarous, Shrikanth S. NarayananInterspeech 2012 · DOI

Abstract

This study utilizes real-time Magnetic Resonance Imaging (rtMRI) to investigate the articulatory characteristics of "emphatic" consonants in Lebanese Arabic. By analyzing the midsagittal plane of the vocal tract, the researchers quantify the secondary constriction in the upper pharynx associated with emphasis. The findings challenge traditional binary phonological models, demonstrating that emphasis is a gradient physical process. Furthermore, the study provides evidence that emphasis spreads to neighboring segments, even through palatal sounds previously hypothesized to act as blockers.

Introduction

In Semitic languages like Arabic, "emphatic" consonants (e.g., /T/, /S/) are distinguished from their "plain" counterparts (e.g., /t/, /s/) by a secondary constriction deep in the vocal tract. Historically, observing this "black box" of the human throat has been challenging. Traditional methods such as X-rays pose radiation risks, ultrasound is limited by bone structures like the hyoid, and endoscopy is highly invasive. Consequently, debates persist regarding the exact location of these constrictions and the nature of "emphasis spread"—the process by which the emphatic quality affects adjacent vowels and consonants. Some theories suggest that certain sounds, such as the palatal /j/, should block this spread, but empirical evidence has been scarce.

Methodology

The researchers employed real-time MRI (rtMRI) to capture high-speed imaging of the vocal tract at 12.5 frames per second, reconstructed to 23.18 fps using a sliding window technique. This non-invasive technique allows for a comprehensive view of the midsagittal plane from the lips down to the glottis.

The study focused on a 28-year-old female native speaker of Lebanese Arabic. Stimuli included minimal pairs (e.g., tamman vs. Tamman) to isolate emphatic effects and specific word structures to test emphasis spread across palatal (/j/) and non-palatal segments. A semi-polar analysis grid was used to measure pharyngeal "aperture" (the distance between the tongue and the pharyngeal wall) across 41 gridlines.

Key Results

The Pharyngeal Constriction

The data reveals that the primary site of emphatic constriction is the upper pharynx, approximately 40mm above the glottis (near the uvula). For the emphatic stop /T/, the mean pharyngeal aperture was measured at 5.69 mm, significantly narrower than the 7.64 mm observed for the plain /t/.

Comparison of vocal tract profiles for plain (tamman) and emphatic (Tamman) segments.

Emphasis as a Gradient Process

Articulatory analysis suggests that emphasis is not a binary "on/off" feature. "Emphaticized" segments—plain sounds affected by a nearby emphatic—exhibit tongue retraction and depression similar to primary emphatics but with less intensity. This "emphatic-lite" profile supports a gradient view of phonological features rather than a rigid categorical one.

Emphasis Spread and the "Blocking" Myth

A significant finding of this study is that the palatal approximant /j/ does not fully block emphasis spread in Lebanese Arabic. While emphasis was slightly attenuated after passing through a /j/ segment, the following sounds still showed significant pharyngeal constriction. Specifically, the aperture in a sound following a palatal segment was 5.39 mm, compared to 4.45 mm for non-palatal contexts and 6.9 mm for true plain segments. This proves that emphasis spread is a phonetically fluid process.

Vocal tract profiles showing emphasis spread across palatal segments (tSayyad) compared to plain (bayyat) and emphatic (bayyaD) contexts.

Conclusion

This pilot study demonstrates the efficacy of rtMRI in quantifying deep vocal tract articulations. The results indicate that emphasis in Lebanese Arabic is a fluid, gradient process that permeates neighboring segments regardless of traditional "blocking" sounds. While the study is limited to a single speaker, it provides a robust framework for future research. Expanding this methodology to a broader range of speakers and dialects will be essential to determine the universality of these articulatory patterns and how they interact with prosodic structures like stress and syllable boundaries.

Summary updated

BibTeX
@inproceedings{israel2012emphatic,
  author = {Assaf Israel and Michael I. Proctor and Louis Goldstein and Khalil Iskarous and Shrikanth S. Narayanan},
  title = {{Emphatic segments and emphasis spread in Lebanese Arabic: a Real-time Magnetic Resonance Imaging Study}},
  booktitle = {Interspeech},
  year = {2012},
  doi = {10.21437/interspeech.2012-581},
}
Figure