Infants rapidly learn word-referent mappings through cross-situational statistical learning. This study shows that 12- and 14-month-old infants can resolve referential ambiguity by accumulating statistical evidence across multiple ambiguous word-scene pairings, rather than relying on a single trial. The research challenges the traditional view that infants use attentional, social, linguistic, or representational constraints to resolve ambiguity. Instead, it suggests that infants use statistical learning to identify consistent patterns across multiple trials, enabling them to learn word-referent mappings even in highly ambiguous contexts. The study involved 28 12-month-old and 27 14-month-old infants who were exposed to 6 word-referent pairs in individually ambiguous trials. Each trial presented two words and two objects without clear associations. Over 30 training trials, infants were exposed to various pairings, allowing them to accumulate statistical evidence. In test trials, infants were presented with a single word and two objects, one being the correct referent and the other a distractor. The results showed that infants looked longer at the correct referent than the distractor, indicating they had learned the word-referent mappings. The findings suggest that infants use statistical learning to identify regularities in word-scene pairings, even when individual trials are ambiguous. This process allows infants to learn word-referent mappings efficiently, even in complex and ambiguous environments. The study also highlights the importance of statistical learning in early word acquisition, showing that infants can learn multiple word-referent pairs through cross-situational statistical learning. The results support the idea that statistical learning mechanisms are crucial for early lexical development, enabling infants to overcome referential ambiguity and learn words effectively.Infants rapidly learn word-referent mappings through cross-situational statistical learning. This study shows that 12- and 14-month-old infants can resolve referential ambiguity by accumulating statistical evidence across multiple ambiguous word-scene pairings, rather than relying on a single trial. The research challenges the traditional view that infants use attentional, social, linguistic, or representational constraints to resolve ambiguity. Instead, it suggests that infants use statistical learning to identify consistent patterns across multiple trials, enabling them to learn word-referent mappings even in highly ambiguous contexts. The study involved 28 12-month-old and 27 14-month-old infants who were exposed to 6 word-referent pairs in individually ambiguous trials. Each trial presented two words and two objects without clear associations. Over 30 training trials, infants were exposed to various pairings, allowing them to accumulate statistical evidence. In test trials, infants were presented with a single word and two objects, one being the correct referent and the other a distractor. The results showed that infants looked longer at the correct referent than the distractor, indicating they had learned the word-referent mappings. The findings suggest that infants use statistical learning to identify regularities in word-scene pairings, even when individual trials are ambiguous. This process allows infants to learn word-referent mappings efficiently, even in complex and ambiguous environments. The study also highlights the importance of statistical learning in early word acquisition, showing that infants can learn multiple word-referent pairs through cross-situational statistical learning. The results support the idea that statistical learning mechanisms are crucial for early lexical development, enabling infants to overcome referential ambiguity and learn words effectively.