Shared book reading is widely acknowledged for its positive impact on language development, as it exposes children to complex linguistic structures not typically encountered in daily conversation. However, the mechanisms through which shared reading supports language acquisition remain less well understood. This study investigates the effects of speech complexity and gesture use on children's real-time word learning from books. Using a dual head-mounted eye-tracking paradigm, we assessed gaze dynamics in 18- to 24-month-old children during naturalistic book reading with their parents. Our findings indicate that while parental speech is rich in linguistic diversity, children at this age exhibit a preference for simpler sentence structures. Simpler sentences and imperatives, particularly when paired with child gestures, appear to capture children's attention most effectively. This study emphasizes the interplay between speech complexity, gesture, visual attention, and word learning, demonstrating that multimodal input plays a critical role in facilitating language acquisition.