We set forth to show that lexical connectivity plays a role in understanding early word learning. By considering words that are learned in temporal proximity to one another to be related, we are able to better predict the words next learned by toddlers. We build conditional probability models based on data from the growing vocabularies of 77 toddlers, followed longitudinally for a year. This type of conditional probability model outperforms the current norms based on baseline probabilities of learning given age alone. This is a first step to capturing the interaction between a child’s productive vocabulary and their learning environment in order to understand what words a child might learn next. We also test different types of variants of this conditional probability and find that not only is there information in words that are learned in proximity to one another but that it matters how models integrate this information. The application of this work may provide better cognitive models of acquisition and perhaps allow us to detect children at risk for enduring language difficulties earlier and more accurately.