Is alphabet biasing in bialphabetical word perception automatic and prelexical?

Area of Science:

• Psycholinguistics
• Cognitive Psychology
• Computational Linguistics

Background:

• Phonological ambiguity in Serbo-Croatian words leads to slower and less accurate identification compared to unambiguous words.
• Visual word recognition models need to account for language-specific features, such as script variations.

Purpose of the Study:

• To investigate methods for reducing phonological ambiguity effects in Serbo-Croatian word recognition.
• To examine the role of alphabet-specific contexts in visual word processing.

Main Methods:

• Five experiments were conducted using masked nonword contexts preceding target words.
• Stimuli involved phonologically ambiguous and unique Serbo-Croatian words presented in Roman and Cyrillic scripts.
• Forward and backward masking techniques were employed to manipulate stimulus presentation.

Main Results:

• Alphabet-specific nonword contexts reduced ambiguity effects when congruent with target words.
• Performance improved when target words and masking contexts shared the same alphabet (Roman or Cyrillic).
• Both forward and backward masking showed benefits from alphabetical congruence.

Conclusions:

• Prelexical, automatic processes are influenced by the alphabetical congruence of context and target words.
• A network model of visual word recognition can explain these findings in Serbo-Croatian.
• Script-specific processing plays a crucial role in mitigating phonological ambiguity.