Re: replacement rates

--- In Nostratica@yahoogroups.com, "H.M. Hubey" <hubeyh@...> wrote:

> I am not sure I understand. I am only discussing creating a

semantics

> metric for a short
> list of words. I only suggest that the short list size be a power

of 2

> so that we can use
> binary/boolean concepts. Any list has to be finite in order to be
> useful. We cannot
> handle infinity.
>
> Geraldine Reinhardt wrote:
>
> > Sorry. I'm not clear as to how you will deliniate your "Hubey"

list.

> > If there is no final number, then the list should exist
> > for infinity? Right?
> >
> > Gerry
> >
> > ----- Original Message -----
> > From: H.M. Hubey <mailto:hubeyh@...>
> > To: Nostratica@yahoogroups.com

<mailto:Nostratica@yahoogroups.com>

> > Sent: Tuesday, February 04, 2003 4:48 PM
> > Subject: Re: [Nostratica] Re: replacement rates
> >
> > No final number. This will work for something like the Swadesh
> > list. There are other
> > ways to get semantic distances from dictionaries. I would be
> > interesting to compare
> > them after they have been worked out.
> >
> > Geraldine Reinhardt wrote:
> >
> >> Interesting. Does that leave the final number at infinity

(in

> >> the power of 2)?
> >>
> >> Gerry
> >>
> >> ----- Original Message -----
> >> From: H.M. Hubey <mailto:hubeyh@...>
> >> To: Nostratica@yahoogroups.com
> >> <mailto:Nostratica@yahoogroups.com>
> >> Sent: Tuesday, February 04, 2003 9:51 AM
> >> Subject: Re: [Nostratica] Re: replacement rates
> >>
> >> IT would have to be in powers of 2. I got started on 32
> >> already. We can double it
> >> to 64, then 128. We'd have to add 28 more to Swadesh

100.

> >> Then we can double to
> >> 256, and we only need 28 more for that.
> >>
> >> Gerry wrote:
> >>
> >>> Most interesting. Yet based on Boolean ideas, how many
> >>> words would be contained in renovating the Swadesh list?
> >>>
> >>> Gerry
> >>>
> >>> ----- Original Message -----
> >>> From: H.M. Hubey <mailto:hubeyh@...>
> >>> To: Nostratica@yahoogroups.com
> >>> <mailto:Nostratica@yahoogroups.com>
> >>> Sent: Tuesday, February 04, 2003 6:01 AM
> >>> Subject: Re: [Nostratica] Re: replacement rates
> >>>
> >>> There is a shorter way. We can start with a subset

> >>> the Swadesh list. Yakhontov
> >>> already has a shorter version which has 35 words. A
> >>> simple metric should probably
> >>> be based on Boolean ideas. 32 is a good number

since it

> >>> is a power of 2. I already
> >>> got started but haven't had time to pay attention

to it.

> >>> If anyone is interested, we can
> >>> collaborate.
> >>>
> >>> tgpedersen wrote:
> >>>
> >>>> --- In Nostratica@yahoogroups.com, "H.M. Hubey"
> >>>> <hubeyh@...> wrote:
> >>>> >
> >>>> >
> >>>> >
> >>>> >
> >>>> > I think what linguistics needs is a "semantic
> >>>> metric". I know how
> >>>> to
> >>>> > derive at least two different
> >>>> > ones from data but need time and money :-) I

think

> >>>> all linguists
> >>>> should
> >>>> > appreciate the need for
> >>>> > semantic metrics. It does not matter how many.

That

> >>>> could always be
> >>>> > worked out, fixed,
> >>>> > repaired, improved etc. The trick is to get

going.

> >>>> >
> >>>> That would require us to know some canonical
> >>>> development of a
> >>>> canonical set of concepts. That is a tall order.
> >>>>
> >>>> Torsten

If you are serious about your semantics metrics, I recommend you read
some of Doug Lenat's articles on his program "Eurisko", a program
that built a conceptual world from axioms of set theory until it ran
into computational difficulties constructing prime number theory. It
will give you a perspective on the problems at hand. They can be
found in the journal "Artifical Intelligence" some time in the early
eighties.

Torsten