Re: what is a loglang?

Me:
> In another less redundant sense, it is strictly a 'logic(al) language',
> & I interpret that as being a language whose grammatical rules specify
> an explicit mapping from surface (phonological) forms to logical
> forms (propositions), with 'logic' understood as propositional
> and predicate logic or some analogue of it.
>
> If someone were to ask "Why fetishize logic by singling it out so",
> I would respond that it is an especially important ingredient of
> the syntagmatics of semantics, and that it could reasonably be said
> that by 'logic' what we really mean is the 'syntax of semantics'.
Mark Line:
> Actually, my next question would be "Why fetishize propositional and
> predicate logic and its analogues?". It seems like there would be a lot
> less impedance matching to do if a logic were chosen that was more like
> language. (No natlang has variables. FOPC quantification does not map
> straightforwardly onto any identifiable natlang phenomenon. So choose a
> logic that gets by without variables and that does quantification like
> natlangs do.)
Well, by 'analogue' I was thinking of anything that might be closer
to the form of everyday thought, so I'd be all for a logic that is
more like thought -- as determined by our ability to discriminate
among different ideas and propositions. I'd be less happy to choose
a logic that closely matches the surface forms of language, since
they are so prone to ambiguity (e.g. "I didn't see everybody",
ambiguous between "I saw nobody" and "Not everybody was seen by me"),
yet people are perfectly capable of conceptually distinguishing the
alternate readings.

I'd be very interested in a formalism that could do without variables
yet could unambiguously represent conceptually discrete propositions.

Wm Drewery:
> I'm not sure I would agree with the way logic is being
> discussed here
>   It seems to me logic is more basic than semantics.
> Any form of intelligence would have exactly the same
> system of logic we do (Boolean), although it's
> formulation may be different, but the semantic gap may
> be huge.. especially if there are major differences in
> sensory perception
This is a question on which in private thought I am wont
to expend a great deal of ill-informed musing. While I
would not rule out the existence of a form of alien
intelligence operating with a logic mutually unintelligble
with our human logic (and therefore essentially beyond
our human ken), I would put my philosophical money
on human logic not being unique to the species -- i.e. on
it being transhuman. This is because it is so hard to
conceive of any alternative to human logic: it is easy
to believe that many human thoughts are transhuman, and
very difficult to think of any radically different way
in which such thoughts could be structured.[*] Yet I am
also of the view that natural language is (or will turn
out to be) speakable logic. Hence I am of the view that
part of semantics (and hence part of language) is
transhuman. In our current state of knowledge, all this
is pretty much entirely speculative, though.

[*] I recall reading an argument by Jackendoff (I think,
but I don't remember where) arguing along these lines
that chimpanzees must necessarily conceptualize
predicates and arguments.

Mark Line again:
> 1. Logic is an attempt to describe (and usually formalize) the
> way(s) in which humans reason; semantics is an attempt to
> describe (and usually formalize) the way(s) in which humans use
> language to mean stuff. Logic is more basic than semantics only
> if reasoning is more basic than language -- me, I think they go
> hand-in-hand.
That is an uncontroversial definition of what logic is, but I
think that I am not atypical of loglangers in viewing logic
-- for the purposes of conlanging -- not as an aid to reasoning
but rather as a perspicuous formalism for representing thoughts.
The more perspicuously and comprehensively it represents thoughts,
the more satisfactory it is.

BP:
> In my experience the preoccupation with propositional and
> predicate logic and its analogues in certain linguistic
> is ultimately ancillary to the desire to process natural
> language with computers.
I think your experience is not representative of the field as
a whole. Many philosophers and linguisticians preoccupied
with logic have no interest at all in NLP, and the preoccupation
with logic is just a preoccupation with the structure of
propositional thoughts.

--And.

Reply