Re: Ambiguity and the interchangeability of parses

Bethan Tovey-Walsh writes:

> Hello, all,

> There was some discussion in Tuesday's meeting of ambiguous parses and
> whether it was useful to report them. In particular, I think someone
> suggested that it didn’t matter which parse the user receives in the
> case that there’s more than one valid parse. I don’t think I did a
> good job of explaining why I think that’s wrong, and I want to put my
> view about this on record.

For what it's worth ...

I agree pretty thoroughly with what you say:  ambiguity can produce
parses whose differences produce only a shrug, or parses whose
differences are important semantically.  I took the remark made on the
call to be short-hand for:  if the user cares about which parse is
produced, then the user should fix the grammar so that the sentence is
not ambiguous.  (And if the user has not fixed the grammar, perhaps it's
because the user does not in fact care.  Perhaps the user knows that
there is some hard-to-eliminate ambiguity in the treatment of whitespace
in some particular part of the expression language, but also knows that
it does not matter for the downstream consumer.)

The 'if you care, then make the grammar unambiguous' position is
probably more plausible when it comes to artificial notations than when
it comes to natural language, as you illustrate well.

(Usage note: ambiguity is not classed as an issue of grammar hygiene, in
part because unlike hygiene issues ambiguity cannot be eliminated with a
programmatic cleanup procedure.  It cannot even be reliably decided
whether a given grammar is or is not ambiguous.)

> In any case, I just wanted to note this as some background to explain
> why I think that the ability to report ambiguity, and to return all
> possible parses, may be very important. It also, I hope, explains why
> I’m very opposed to the idea that the parses returned by an ambiguous
> grammar are always essentially equivalent to each other.

But I don't think anyone was arguing that processors should no longer be
*allowed* to flag ambiguous sentences; I think the strongest position
anyone has taken is that it should not be *required*. 


-- 
C. M. Sperberg-McQueen
Black Mesa Technologies LLC
http://blackmesatech.com

Received on Sunday, 27 March 2022 02:49:52 UTC