NL-KR-REQUEST@CS.ROCHESTER.EDU (12/30/87)
NL-KR Digest (12/29/87 20:10:58) Volume 3 Number 65
Today's Topics:
Re: Language Learning
Re: natural language examiners
Re: semantics of "unless"
Submissions: NL-KR@CS.ROCHESTER.EDU
Requests, policy: NL-KR-REQUEST@CS.ROCHESTER.EDU
----------------------------------------------------------------------
Date: Wed, 16 Dec 87 08:28 EST
From: sas@bfly-vax.bbn.com
Subject: A few more bits on Language Learning
1) Earlier this year there was an article in Science about quail
having a natural ability to distinguish certain phonemes, while having
a lot of trouble distinguishing others. I forget the phonemes, but I
think the confusion was with p and b sounds. I mentioned this to a
linguist friend who pointed out that a similar study had been done
years before, except with chickens. This sounds like birds have an
innate ability to learn human languages, but is probably just an
artifact of the common elements of the mammalian and avian ears.
(Insect and reptilian ears are a bit different, but for all I know
salamanders can learn a few phonemes, too).
2) Has anyone seen any of those neat books actors use when they want
to affect an accent? Look in the drama and theatre section of your
college bookstore.
3) A friend of mine tells me that her first language was Thai which
she learned from her amah when she was growing up in Thailand, but
she stopped speaking it a long time ago and has forgotten it all. I'm
sure I had a great French accent when I was a child, but then again I
remember when a coffee table was as big as a candy store.
Seth
------------------------------
Date: Wed, 16 Dec 87 12:37 EST
From: Brady@UDEL.EDU
Subject: Language learning
Those of you debating how language is learned might be
interested in referring to Dr Phillip Lieberman's book
"The Biology and Evolution of Language" (1984, Harvard
University Press). Dr Lieberman is professor of linguistics
and cognitive science at Brown University. In his book
he explains language understanding and production (including
the existance of a "critical period") from
biological and evolutionary perspectives.
------------------------------
Date: Wed, 16 Dec 87 17:40 EST
From: Paul Placeway <paul@ptero.cis.ohio-state.edu>
Subject: Re: Language Learning
In article <3136@bcsaic.UUCP> rwojcik@bcsaic.UUCP (Rick Wojcik) writes:
< In article <430@minya.UUCP> jc@minya.UUCP (John Chambers) writes:
< >...
< >Sure, people will *say* that they would like to speak like a native,
< >just like they *say* that they'd like to learn to play the piano or
< >lose 20 pounds or clean up the basement. But really, the problem just
< >might be that they in fact don't have any real motive to invest the
< >time that it would take.
< >...
< There are two implications of your argument that you ought to consider.
< First of all, it implies that children who lack motivation should not
< learn to speak as well as those who are diligent and highly motivated.
< We find that attitudes do affect learning in children, but I know of no
< evidence that it makes for poorer language production and comprehension.
< This is because children acquire language independently of training and
< rote learning.
Ah, but you are assuming that there is such a thing as a child who has
no motivation to learn to understand and speak a language. The only
example of such I can think of is a learning disabled child, who
generally _don't_ learn to speak as well as a "normal" child. (I'm
not doing a "I can't think of it, so it must not exist", but I would
need to see several _solid_ examples to even consider the possibility.)
< The second implication of your argument is that highly
< motivated adults can achieve flawless speech in a new language.
I disagree. I think that the implication is that the the prospective
learner assumes that (a) it is possible, and (b) that it will require
a _lot_ of effort. Note that the assumption of the learner is quite
independent of the facts: the learner can assume that some action is
possible even if it is not, choose not to try it, and never be shown
that they were wrong. Without any evidence to the contrary, they can
_still_ hold that belief.
< You seem to have my number. I like to *say* I'll do a lot of things.
< Are you trying to tell us that children aren't like this too? Maybe
< they are all enthusiastic about language acquisition before puberty, but
< not about things like cleaning up the room or learning to play the
< trumpet :-]. Then, after puberty, they suddenly get lazy about learning
< languages, too. So, why the sudden change in attitude towards language?
< And why is everyone losing interest at the same time?
Earler you have said (quite correctly) that language aquisition and
learning a language (as in in school) are different; why are you
assuming that they are the same here? Many very young children think
that learning new words is fun, but that grammar class is boring.
Many kids like to read for fun, but hate to for school. I don't think
that kids have any different attitude toward _learning_ a language as
adults do.
I think that the real question is: why don't adults _aquire_ new
languages in the same way that children do? (which is what you seem
to be really asking)
-- Paul
------------------------------
Date: Fri, 18 Dec 87 12:12 EST
From: John Pantone <sdcc6!calmasd!jnp@sdcsvax.ucsd.edu>
Subject: Re: Language Learning (anecdotes)
(Douglas Moreland) writes:
>Though news announcers are subject to the firing and hiring whims of their
>bosses, ... The language of news announcers thus becomes a standard dialect
>...
I agree, in general Doug, but there are some notable exceptions: Bill Moyers
has a very distinct(ive) southern accent (sorry - can't place it better than
that - I am not familiar with the various southern accents) and in addition
uses phrases which are, at least to me, characteristic of southern speakers.
I wonder if he writes/wrote most of his own editorials?
My guess is that the copywriters on major news programs assiduously avoid
regional dialect, and opt for a "bland" non-colloquial form of speach.
There is, to my ear, a "typical" news-show dialect, somewhat different than
that I've ever heard anyone speak on the street - but very understandable
and un-ambiguous.
Example?
"The President announced, today, that there will be ..."
"<reporter>, our reporter on the scene, reports..." (not has reported, or
will report)
etc.
--
These opinions are solely mine and in no way reflect those of my employer.
John M. Pantone @ GE/Calma R&D, 9805 Scranton Rd., San Diego, CA 92121
...{ucbvax|decvax}!sdcsvax!calmasd!jnp jnp@calmasd.GE.COM GEnie: J.PANTONE
------------------------------
Date: Fri, 18 Dec 87 06:27 EST
From: Charles Lambert <mcvax!ukc!stc!datlog!dlhpedg!cl@uunet.uu.net>
Subject: Re: natural language examiners
In article <733@csinn.UUCP> grossi@csinn.uucp (Thomas Grossi) writes:
>(for example, if someone steps on your foot and apologizes -- "Excuse me" --
>an appropriate response in English would be "certainly" whereas in French
Appropriate response varies within a language as well as between languages.
For instance, the above apology would be appropriate in America, but in
England "Excuse me" is often interpreted as "Get out of my way"; "Pardon me"
or "I beg you pardon" would be safer (again, variant from American usage).
-----------
Charles Lambert
------------------------------
Date: Mon, 21 Dec 87 13:36 EST
From: berke@CS.UCLA.EDU
Subject: (A unless B) <-> (B implies not A) [from Berke]
(Note: The truth table and the original posting seem to have reversed
the role of A and B. The reversal below is intentional.)
A clear translation of (A unless B) into propositional calculus
may be made by noting the similarity between (A unless B) and
(A if B).
(A if B) is commonly translated as (B implies A).
(A unless B) is commonly translated as (B implies not A),
that is, (not-A if B).
That (A unless B) seems to be "saying something" about A,
ignores the binary-relational nature of connectives in PC.
For example, if we are given (B implies A) and (not B), we cannot
infer either (A) or (not A). We can see this by transforming
(B implies A) to its equivalent form ((not B) or A).
If we are given (B implies not A) and (not B), we similarly cannot
infer either (A) or (not A).
More complex translations of (A unless B) may be made, but
the same is true of (A if B). Beware, some attempts to
clarify these distinctions result in Modal Logic.
Calling other translations "more complex" is relative to my choice
of implication and negation as primitive. If you desire a sense
of symmetry in your "unless," and if you're really into a particular
kind of simplicity, try "not both," i.e., (A nand B) which may be
taken as the only primitive connective in a formulation of the
propositional calculus that has only that one primitive
connective, one rule of inference, and one axiom.
The original posting stated that Lang was asked to prove "A unless B."
All the responses (including mine above) are quite general so, if
the person asking the question had some clear intended meaning for
'unless', they are all irrelevant. I suppose that A and B were
complex terms in the original problem statement, otherwise we would
have difficulty proving any of our translations of 'unless'. I also
suppose there was some intended proof system. I would suppose that
there were some clearly intended translation for "unless," but
there wasn't or we wouldn't be discussing this. Problems of the
sort "translate this into PC" are often equivocal (can be translated
in more than one way) precisely to get the student to think about
the options and to emphasize formalization is a matter of choice.
Since the problem asked for a "proof" rather than a translation,
I would conclude that it presupposed an unequivocal (univocal)
translation for "unless," which is to be found earlier in the
chapter. If it is not to be found earlier in the chapter, yet
still requests a proof, the problem statement is faulty.
Respectfully,
Peter Berke
------------------------------
Date: Mon, 21 Dec 87 07:10 EST
From: Christian Ronse <ronse@prlb2.UUCP>
Subject: Re: semantics of "unless"
In article <149@piring.cwi.nl>, varol@cwi.nl (Varol Akman) writes:
> lang@bigburd.PRC.Unisys.COM (Michel Lang) writes:
> >Does anybody have any thoughts about the truth-functional meaning
> >of the connective "unless"? What I'm getting at is the following:
> >I recently came across a problem that asked me to prove that
> >
> > P unless Q
> >
[etc.]
> >I have since decided in favor of P <==> ~Q, but the number
> >of disagreeing answers I received from colleagues made me think
> >it might be worth posting. Any ideas?
> >
> P unless Q
>
> can be understood as (from a logical viewpoint)
>
> P should be implied by ~Q
I agree. ``P unless Q'' means that P is true, with possible exception only if
Q is true. It is thus logically equivalent to ``~P=>Q'', in other words ``P OR
Q''. The difference between OR and unless is not logical, but semantical:
``unless'' means that the case where P is false is somewhat an exception. Cfr.
the driving code, and the two priority signs:
^ ^
/ \ / \
/ ^ \ / \
/ -|- \ \ /
/ | \ \ /
--------- v
Both mean you have priority, but the first one generally appears as an
exception, while the second one often comes in a sequence.
Christian Ronse maldoror@prlb2.UUCP
{uunet|philabs|mcvax|...}!prlb2!{maldoror|ronse}
------------------------------
Date: Mon, 21 Dec 87 11:58 EST
From: Rick Wojcik <rwojcik@bcsaic.UUCP>
Subject: Re: semantics of "unless"
Those interested in this subject might wish to read Michael Geis' "If
and Unless" in B. Kachru et al., eds. Issues in Linguistics. Univ. of
Illinois Press. 1973. Geis is on the faculty of the linguistics dept.
at Ohio State. Unfortunately, I don't think he has a net address.
===========
Rick Wojcik rwojcik@boeing.com
------------------------------
Date: Tue, 22 Dec 87 11:31 EST
From: Craig_Presson@RIVERBOAT.CEO.DG.COM
Subject: Re: semantics of "unless"
From: Craig%Riverboat.ceo.dg.com@RELAY.CS.NET
Date: Tues, 22 Dec 1987, 11:17 EST
From: Craig Presson <Craig%Riverboat.ceo.dg.com@RELAY.CS.NET>
Subject: Semantics of "unless"
In the posting (from Michel Lange, Vol 3, no. 64) about the semantics of
"unless", there are really two questions: first, does the English construct
"P, unless Q." _ideally_ match the PC expression "~Q -> P" (equiv. P or Q)
or the expression "~Q <-> P" (equiv. P xor Q); second, given that we
choose correctly the _ideal_ translation, how much trouble are we in when
_real_ sentences come along where the writer is ignoring the difference,
exploiting the difference, or just leaving it to "context"? There is a
similar, and in fact, worse problem around the translation of "or".
Sometimes "or" in English corresponds to "or" in PC, and sometimes it
corresponds to "xor". When the difference really matters (e.g., in writing
functional specs for software systems), some writers ignore the ambiguity,
some try to make it clear from context, and some use extra verbiage such as
"either ... or" for exclusive or, or the grammatically disgusting "and/or"
for inclusive or.
The issue with "unless" happens to revolve around the same case (whether
the valuation P=t, Q=t satisfies the sentence or falsifies it) as the
ambiguity of "or". I looked at the possible truth tables, cheated and did
some English examples, and concluded that I like the "~Q <-> P" translation
better "ideally". One can state a spectrum of examples:
1) The journal you want is somewhere in my messy office, unless I left it
at home.
2) I intend to go to the meeting, unless I am sick.
3) The US will continue SDI, unless the USSR leaves Afghanistan.
In example 1, the intention to "exclude the middle" is fairly clear,
_unless_ (:-)) the speaker's office is in his home. In example 2, there is
an issue of temporality: a valuation is possible only at some indefinite
future time when it is necessary to evaluate the speaker's health, at which
time, if she is sick, and forms the intention of going to the meeting
ANYWAY, I claim she has falsified her earlier statement, so the "<->"
semantics is still supported; and I quite agree that example 3 leaves one
to judge from some unspecified context whether the speaker has "<->" or
"->" in mind. (So, what did they REALLY say at the summit? ;-)
------------------------------
Date: Wed, 23 Dec 87 10:32 EST
From: 55551-K.J.Anderson <krista@ihlpa.ATT.COM>
Subject: Re: semantics of "unless"
In article <531@udccvax1.acs.udel.EDU>, chiefdan@vax1.acs.udel.EDU (ROTH) writes:
>
> lang@bigburd.PRC.Unisys.COM (Michel Lang) writes:
> >>Does anybody have any thoughts about the truth-functional meaning
> >>of the connective "unless"? What I'm getting at is the following:
> >>I recently came across a problem that asked me to prove that
>
> >> P unless Q
Well, there are several conclusions one can reach:
1. P xor Q
2. ~Q implies P
3. Q implies ~P
and others. But, the English word "unless" has an additional
connotation. The statement above implies that P is more probable
than Q, that P is the rule and Q is the exception to the rule.
I don't think there's any way to translate that into a predicate
calculus, and you could draw mathematical conclusions only if you
knew the probabilities of P and Q.
K. J. Anderson
--
ihnp4!ihlpa!krista
------------------------------
Date: Thu, 24 Dec 87 09:46 EST
From: Murray Watt <murrayw@utai.UUCP>
Subject: Re: semantics of "unless"
In article <6616@ihlpa.ATT.COM> krista@ihlpa.ATT.COM (55551-K.J.Anderson) writes:
>
>Well, there are several conclusions one can reach:
>1. P xor Q
>2. ~Q implies P
>3. Q implies ~P
>
>and others. But, the English word "unless" has an additional
>connotation. The statement above implies that P is more probable
>than Q, that P is the rule and Q is the exception to the rule.
>
>I don't think there's any way to translate that into a predicate
^^^^^^^
>calculus, and you could draw mathematical conclusions only if you
>knew the probabilities of P and Q.
>
Define a set of binary predicates, {unless#1, unless#2, unless#3,...}
over propositions.
unless#1(P,Q) iff P xor Q and the precentage of worlds and times that
P is true is greater than the precentage of
worlds and times that Q is true by at least
some precentage Y1.
unless#2(P,Q) iff ~Q implies P and the precentage of worlds and times that
P is true is greater than the precentage of
worlds and times that Q is true by at least
some precentage Y2.
unless#1(P,Q) iff P implies ~Q and the precentage of worlds and times that
P is true is greater than the precentage of
worlds and times that Q is true by at least
some precentage Y3.
The Ys depend on the particular implementation or person.
Yes, you have to know the probabilities, but that is true of many logical
operators (eg. necessary, probably, possibly,...).
Murray Watt was here!
------------------------------
End of NL-KR Digest
*******************