This reminds me … of something Jerry Norman was wont to say, i.e., that there were three good criteria for identifying Mandarin and deciding how old the family is. These are the concurrent presence of the third person pronoun tā, the negative bù, and the subordinative particle de/di. Jerry called languages of this type “Tabudish”, and he sometimes used this name for them in correspondence with me.
Other commenters on the Shanghainese post, especially Tsu-Lin Mei, gave additional, precise criteria for distinguishing Mandarin from Wu, which led them to conclude that the roots of Mandarin go back before the Tang Dynasty (618-907) to the Six Dynasties period (220/222-589). I should note that both Tsu-Lin and South were close associates of Jerry Norman, and the three of them together have made remarkable contributions to the understanding of the early rise of Mandarin.
As for how much further the beginnings of Mandarin per se might be pushed, I wouldn't care to venture, but I have little doubt that the split between Literary / Classical and Vernacular Sinitic goes back to B.C. times. Although three millennia of literary redaction have left precious little evidence of the vernacular before the Tang period when Buddhism began to legitimize its written form (see Victor H. Mair, "Buddhism and the Rise of the Written Vernacular: The Making of National Languages,” Journal of Asian Studies, 53.3 [August, 1994], 707-751), we do find occasional bits and pieces of the vernacular that have managed to slip through the grasp of the literary editors of the textual tradition. Even more exciting is the discovery of archeologically recovered texts which help to document the existence of the vernacular during the B.C. era.
One of the clearest indications of Vernacular Sinitic is the use of shì 是 as the copulative rather than as the demonstrative pronoun as in Literary / Classical. Rare examples of this usage have been showing up in recently unearthed texts. About six or seven years ago, Jeff Rice wrote a brilliant paper in which he showed how shì 是 evolved from being used for the Literary / Classical demonstrative into the copulative verb in Vernacular. At the same time, he documented the shift from the Classical form X Y yě 也 to Vernacular X shì 是 Y for equational sentences ("X is Y"). Unfortunately, although from time to time I've nudged Jeff to publish that paper, it's still moldering is some drawer. Maybe now that he's finished his dissertation on medieval historiography, perhaps I'll be able to persuade him to publish the paper on shì 是 and yě 也 before another six or seven years pass.