In an earlier post (March 24, 2012), I floated an ad hoc typology of the contemporary use among younger American speakers (under the age of 25?) of the disfluent filler or discourse marker like, as follows:

(1) TICASTIC: for many speakers, the word is a verbal tic (whence the nonce adjective “ticastic”), replacing “you know” and its congeners, and having no other function than as a meaningless filler;
(2) PHATIC (perhaps as a sub-species of the ticastic): keeping the channel of communication open, sometimes for no other reason than to forestall a response from one’s interlocutor(s);
(3) QUOTATIVE: as a prefatory marker before the report of someone else’s utterance(s) or inner speech;
(4) APPROXIMATIVE: as a means of qualifying the extent or validity of the word or phrase immediately following, including its literal meaning;
(5) ANAESTHETIC: as a way of deflecting the assertory force of anything following, usually as an apotropaism.

Prompted by new specimens of raw speech overheard viva voce into thinking further about the distribution of approximative and quotative like, I now suspect that the latter may be derivative of the former. The logic behind this relation resides in the implied judgment that no report of direct or indirect speech can ever be precise because only the speech act itself––and not its retelling––can ever authentically stand for itself. By this logic, no statement of anything that contains figurative expressions can ever be considered verisimilar. With respect to the use of the word like, this would then have the advantage of accounting as well for the currently ticastic British qualifying phrase (pre- or post-posed), if you like.