The double deficit hypothesis states that naming speed problems represent a second core deficit in dyslexia independent from a phonological deficit. The current study investigated the main assumptions of this hypothesis in a large sample of well-diagnosed dyslexics. The three main findings were that (a) naming speed was consistently related only to reading speed; (b) phonological processing speed and naming speed loaded on the same factor, and this factor contributed strongly to reading speed; and (c) although general processing speed was involved in speeded naming of visual items, it did not explain the relationship between naming speed and reading speed. The results do not provide support for the existence of a second independent core naming deficit in dyslexia and indicate that speeded naming tasks are mainly phonological processing speed tasks with an important addition: fast cross-modal matching of visual symbols and phonological codes. (Contains 1 figure and 8 tables.)