Stages of Sight Translation

Sight translation—reading a document in one language while speaking it aloud in another—is one of the most demanding tasks a bilingual mind can perform. How do professional interpreters seamlessly coordinate text reading with instant oral translation?

In a study published in Applied Psycholinguistics, researchers Agnieszka Lijewska, Agnieszka Chmiel, and Albrecht Inhoff used eye tracking to uncover the cognitive stages behind this complex skill.

Eye Tracking the Stages of Translation

The researchers tested 24 professional conference interpreters and 15 interpreting trainees as they sight-translated English sentences into Polish. To evaluate how interpreters handle translation difficulty, the source sentences contained critical nouns manipulated for word frequency (high-frequency vs. low-frequency words).

To capture the subtle temporal dynamics between visual processing and vocal output, the researchers relied on an EyeLink 1000 Plus eye tracker to monitor eye movements as sentences appeared on a monitor. At the same time, speech was recorded with a microphone.

The EyeLink eye tracker enabled the researchers to cleanly separate early reading processes from late reading processes:

1. First-Pass Measures: First fixation duration, gaze duration, and go-past time.
2. Second-Pass Measures: Re-view duration and total viewing duration.

Multi-Stage Model of Translation

The granular gaze data collected by EyeLink provided strong empirical evidence for a multi-stage model of sight translation:

1. First-Pass Reading (Comprehension Stage): Translation expertise had no effect on initial reading times (first fixation or gaze duration). Initial text scanning during sight translation functions just like regular reading for comprehension, regardless of skill level.
2. Second-Pass Re-Viewing (Translation Stage): The actual heavy lifting of translation occurs during re-reading. Here, expert interpreters demonstrated significantly shorter re-view durations than trainees—a professional advantage that was especially pronounced on difficult, low-frequency words.

Translation is a Process

This eye tracking research illustrates that sight translation is not a single, monolithic action. Instead, it unfolds sequentially: interpreters first comprehend the text much like standard silent reading, and only then shift into a specialized, cognitively demanding re-reading phase where the heavy lifting of translation truly happens.

有关眼动追踪如何帮助您的研究的信息，请查看我们的解决方案和产品页面或联系我们。我们很乐意为您提供帮助！