How to extend the anti-luminosity argument to synonymy and meaning?
To what extent is that possible? Williamson is quite confident about it in Knowledge and its Limits but it's really unobvious to me how to construct a soritical series of cases with small changes of meaning. It is conceivable with colour words (with stipulated sharp cut-offs) shifting their boundaries, but otherwise seems hard. He talks about synonyms gradually differing in tone and then in meaning but I can't really imagine that as a quasi-soritical series. Changes in tone seem to me to be qualitative (for example how offensive an expression is seems to me to be measured by its qualitative effects on speakers), and I can't conceptualize a shift in tone that suddenly becomes a shfit in meaning - how could that be indiscriminable/tolerated in the relevant sense?
Besides, in his debate with Boghossian, Williamson himself convincingly argues that, for example, slurs refer to real properties, even if they are employed with fallacious inference-rules (so pace Dummett, an offensive slur used to refer to members of that group, really does refer to the property of being in that group). That makes the example of a shift in tone gradually slipping into a shift in meaning even stranger.
Are applications of anti-luminosity to the philosophy of language discussed in the literature? (Especially their force against Dummett)
I am also unsure about his example of gradually forgetting a mathematical proof as an anti-luminosity argument against Dummettian intuitionism. Couldn't we say that there is a sharp cut-off precisely at the point at which the subject lacks the ability to produce all the steps of the proof?