Understanding words requires infants to not only isolate words from the speech around them and delineate concepts from their world experience, but also to establish which words signify which concepts, in all and only the right set of circumstances. Previous research places the onset of this ability around infants' first birthdays, at which point they have begun to solidify their native language phonology, and have learned a good deal about categories, objects, and people. In this dissertation, I present research that alters this accepted timeline. In Study 1, I find that by 6 months of age, infants demonstrate understanding of around a dozen words for foods and body parts. Around 13-14 months of age, performance increases significantly. In Study 2, I find that for a set of early non-nouns, e.g. uh-oh' and eat', infants do not show understanding until 10 months, but again show a big comprehension boost around 13-14 months. I discuss possible reasons for the onset of noun-comprehension at 6 months, the relative delay in non-noun comprehension, and the performance boost for both word-types around 13-14 months. In Study 3, I replicate and extend Study 1's findings, showing that around 6 months infants also understand food and body-part words when these words are spoken by a new person, but conversely, by 12 months, show poor word comprehension if a single vowel in the word is changed, even when the speaker is highly familiar. Taken together, these results suggest that word learning begins before infants have fully solidified their native language phonology, that certain generalizations about words are available to infants at the outset of word comprehension, and that infants are able to learn words for complex object and event categories before their first birthday. Implications for language acquisition and cognitive development more broadly are discussed.