Diagnosing classic autism doesn't particularly hinge on an absence of speech until age three or later, though; highly idiosyncratic speech, echolalic speech, or non-communicative speech all "count" as a speech delay for the purposes of an autism diagnosis. For example, if you spoke mostly in pre-recorded phrases to communicate your needs; or you copied speech patterns without communicating what you wanted to say; or you had speech that had significant grammar/syntax errors (usually with a lot of neologisms), then that would be sufficiently unusual to be considered a sign of classic autism. Many diagnosed "aspies" should be diagnosed classic autistic because of that distinction; highly unusual or non-communicative speech is classic autism... Really, if you apply the criteria strictly, Asperger's fairly fades into classical autism, and being diagnosed AS gets very rare because anyone who fits AS criteria also fits classic autism criteria, and the rest often don't have "significant impairment" and can't be diagnosed at all... I'm rather glad they're going to merge these; it'll stop people diagnosing Asperger's based mostly on how much you fit the AS stereotype!