@phdthesis{cogprints8831, month = {April}, title = {The Processing of Lexical Sequences}, school = {University of Alberta}, author = {Dr. Cyrus Shaoul}, year = {2012}, keywords = {n-grams, frequency, mutual information, lexical probability}, url = {http://cogprints.org/8831/}, abstract = {Psycholinguistics has traditionally been defined as the study of how we process units of language such as letters, words and sentences. But what about other units? This dissertation concerns itself with short lexical sequences called n- grams, longer than words but shorter than most sentences. N-grams can be phrases (such as the 3-gram "the great divide") or just fragments (such as the 4- gram means "nothing to a"). Words are often thought to be the universal, atomic building block of longer lexical sequences, but n-grams are equally capable of carrying meaning and being combined to create any sentence. Are n-grams more than just the sum of their parts (the sum of their words)? How do language users process n-grams when they are asked to read them or produce them? Using evidence that I have gathered, I will address these and other questions with the goal of better understanding n-gram processing.} }