Word Order in Early Language Development: A Computer-Based Study of Turkish-Learning Infants

Infants begin learning the structure of their native language long before they produce or understand words. From the first months of life, they track how sounds and syllables are organised, detecting recurring patterns that guide them to discover the order typical of their language. They rely on cues such as prosody and frequency, which help them identify where short, frequent elements occur in relation to longer, variable ones. These positional regularities differ across languages, creating distinct word-order typologies: frequent-infrequent (FI) in languages such as English and Italian, and infrequent-frequent (IF) in languages such as Turkish and Japanese. Previous research has primarily focused on FI languages, leaving the mechanisms underlying IF systems less understood. Current study aimed to address this gap in the literature by adapting an artificial grammar learning (AGL) paradigm constructed by Gervain et al. (2008). Twenty Turkish-learning infants aged six to twelve months were presented a neutral familiarisation stream followed by eight test sequences built from a simplified AXBY grammar, in which the placement of two high-frequency syllables (“fi”, “ge”) defined the structure of the test stimuli either as frequent-initial (FI) or frequent-final (IF). The task was administered online, providing a controlled but fully remote testing environment. Attention was measured through looking-time, rather than the head-turn preference used in the original study. Results from the paired-samples t-test revealed that infants looked significantly longer (t(19) = 2.54, p = .020) at frequent-final sequences than at frequent-initial ones, indicating that the sensitivity to the distributional patterns of their native word order are present as early as six months of age. These findings extend previous research from function-initial to function-final languages and provide substantial evidence that Turkish-learning infants already attend to structures reflecting their native language typology. Additionally, the study introduces a methodological contribution to artificial grammar learning research by demonstrating their use in online paradigms can produce reliable and valid measures in infant research.