Teaching sight words, particularly the Dolch and Fry lists, is standard practice when teaching beginning and remedial reading. Sight words are singled out for instruction because these high frequency words are necessary for fluent reading; the first 25 words constitute about 33% of the words we read in a text (Fry, 1957), so quick and accurate recognition is important. Many of these words are believed to be phonetically irregular, meaning they do not follow phonic rules, and this has led to the practice of teaching students to read these words “by sight,” by memorizing their visual features and not relying on the sounds of the letters. This paper explores a question: should sight words be taught “by sight?” A review of the current theories and recent gains research has made in understanding how the brain learns to read and process words– including studies by Dehaene (2011); Ehri (2003); Farrell, Osenga, and Hunter (2013b) – suggests that it is time to review and reflect on the current standard practices for sight word instruction. Close evaluation reveals that the majority of sight words adhere to phonetic rules. Based on this knowledge, reading experts recommend teaching students to read sight words through letter-sound relationships using phonics-based instruction (Ehri,1998; Farrell, Osenga, & Hunter, 2013b; Dehaene, 2011; Laurita, 1966; Moats, 1999; Shaywitz, 2003).