I think there is some valid and original thinking in here, that does point to flaws in prevalent conceptions of language, writing, speech and their interrelations.
At certain points Harris (somewhat) recalls Derrida, e.g. 'the sign itself is a product of contextualisation', which one could compare, in context, to 'il n'y a pas de hors-texte'. Which is kind of funny since Derrida is one of many scholars whom Harris is bluntly rude about; in fact, Harris is rather boorish about more or less everyone he quotes and refers to, unless they are members of his particular school, that (possibly fictional?) of 'integrational linguistics' (IL hereinafter).
That's a pity - it's off-putting when a theorist feels the need to put down everyone who isn't signed up to their own way of thinking.
Also, he introduces some IL terminology, piecemeal and without giving any overall presentation of it. This is unhelpful, because it makes it impossible to evaluate much of what he suggests IL has to offer. And while quite a bit of his critique of existing perspectives rings plausible, he does also quite often go in for impressionistic waffle himself (of a kind that he upbraids others for), using familiar but technically undefined expressions and extremely vague propositions such as: "What the semiology of the signature tells us is something about the society responsible for its evolution as a graphic practice." (This is from perhaps the weakest chapter of all, 'On the Dotted Line', about signatures and their ins and outs.)
All the same, I give it quite a high rating because it is a sustained, if hugely uneven, attempt to deconstruct (soz, Roy) the tramlines of language-theoretical orthodoxy.