Towards Truth in Head - Driven Phrase Structure Grammar

Any linguistic paradigm with scientiic aspirations must provide substantial answers to two questions. What does a grammar of a natural language claim of the natural language? And what qualiies as corroborating or confuting evidence for that claim? I have striven for longer than I am willing to casually admit to furnish the HPSG (Head-driven Phrase Structure Grammar) linguistic paradigm of (Pollard and Sag, 1994) with material responses to both questions, and I have felt for some years now that I have nally arrived at partial yet signiicant characterisations of two important notions: truth, what it is for a grammar of a natural language to be true, and veriication, how the truth of a grammar of a natural language is empirically assessed. All HPSGians ought to have an interest in good HPSG characterisations of truth and veriication, since HPSG aaords little of scientiic consequence without them. Here I present only my characterisation of truth, and leave my characterisation of veriication to a later paper. But this is only tting, since my characterisation of truth in very large part informs my characterisation of veriication. I have taken great pains to present my characterisation in a manner that is engaging enough to attract and inform a broad HPSG readership. To this end, I have omitted all mathematical proofs and included only those mathematical deenitions, propositions, theorems and suchlike that are indispensable to understanding the paper. I am sure that the consequent reduction in the intervention of mathematical technicalities between the reader and that which I most wish to convey to the reader will more than compensate for the undoubted mathematical cheapening of the paper. However, the reader should not be lulled into thinking that this paper is easy to read. Far from it. Truth is a profound and subtle matter, and a woolly characterisation of truth ooers merely a utility that is severely curtailed and a comprehensibility that is largely illusory. My characterisation is anything but woolly. It is rigorous enough to avoid omission, imprecision and contradiction. But such rigour demands much not only of the author but also of the reader. This paper is not light bedtime reading. 1. Introduction Characterising truth in the HPSG of (Pollard and Sag, 1994) is far more diicult than most HPSGians realise, and a critical review of 1994 vintage HPSG is necessary before embarking upon a characterisation proper. A notable hallmark of (Pollard and Sag, 1994) …