[sci.lang] Regional accents

Pryor-Louise@cs.yale.edu (Louise Pryor) (05/12/89)

In article <2763@puff.cs.wisc.edu>, brian@cat50.CS.WISC.EDU (Brian Miller) writes:
> 
>    Regional dialects are rapidly evaporating as mass communication becomes
> an integral part of modern society. 


This is just not true. It is still impossible to talk about *the*
British accent, let alone say that British and American accents are
similar.  Just listen to someone from Glasgow, or Birmingham, or
Cornwall, or North Wales, or Norwich, or South London or...  And
that's just in one small island! Alternatively, try to make yourself
understood with an English accent (so-called BBC English) at a filling
station in Atlanta (or New Haven, come to that). I doubt whether
Britons and Americans will ever have the same accent, the vowel sounds
are just too different on the whole.


Louise Pryor
ARPA:   pryor@cs.yale.edu
BITNET: pryor@YALECS.BITNET

goer@sophist.uucp (Richard Goerwitz) (05/12/89)

Pryor-Louise@cs.yale.edu (Louise Pryor) writes:
>I doubt whether
>Britons and Americans will ever have the same accent, the vowel sounds
>are just too different on the whole.
>
>Louise Pryor

What's interesting to me is that, after just a night listening to
this or that British dialect, an American can pretty much understand
everything perfectly.  Actually, a night is being liberal.  In most
cases the adjustment takes no more than a few minutes.  You kinda
lock into their system.

What is it that facilitates this kind of rapid adjustment?  When you
look at the phonetics, the dialects are quite different.  I mean, if
you transcribe a few lines using a broad phonetic notation system, you
get something quite unlike any American dialect.

                                       -Richard L. Goerwitz
                                       goer@sophist.uchicago.edu
                                       rutgers!oddjob!gide!sophist!goer

lammens@sunybcs.uucp (Jo Lammens) (05/13/89)

In article <3193@tank.uchicago.edu> goer@sophist.UUCP (Richard Goerwitz) writes:
>What's interesting to me is that, after just a night listening to
>this or that British dialect, an American can pretty much understand
>everything perfectly.  Actually, a night is being liberal.  In most
>cases the adjustment takes no more than a few minutes.  You kinda
>lock into their system.
>
>What is it that facilitates this kind of rapid adjustment?  When you
>look at the phonetics, the dialects are quite different.  I mean, if
>you transcribe a few lines using a broad phonetic notation system, you
>get something quite unlike any American dialect.

I suspect one quickly learns to map one's own phonological/phonetic
system into that of the person being listened to. The "phonetic mode
of perception" probably has something to do with that (although it is
a debated phenomenon): vowel and other sounds are separated into
classes with sharp class boundaries across an otherwise continuous
spectrum of sound change. So if you learn to place the class
boundaries at another arbitrary point, you suddenly "understand" what
someone is saying, although he may use a rather different sound
system. Of course there is more to it than just the spectral
composition of speech sounds, but the same probably holds mutatis
mutandis for things like temporal aspects (VOT, reductions, ...),
intonation patterns and the like. 

In fact there is no such thing as THE B.E. or A.E.  pronunciation, or
even THE pronunciation for any one dialect. If one acoustically
analyses people's pronunciation of the same words, the differences
tend to be very large even among speakers of the same dialect. Yet no
one really notices. That is one of the reasons that automatic speech
recognition is so difficult: the identification problem for phonemes
is a complex one, but you don't realize it until you start measuring
things or try to build a system that does it. We humans can easily
shift our perceptual category boundaries around without even being
conscious about it.

Jo Lammens

BITNET: lammens@sunybcs.BITNET          Internet:  lammens@cs.Buffalo.EDU
UUCP: ...!{watmath,boulder,decvax,rutgers}!sunybcs!lammens

hollombe@ttidca.TTI.COM (The Polymath) (05/16/89)

In article <60340@yale-celray.yale.UUCP> Pryor-Louise@cs.yale.edu (Louise Pryor) writes:
}In article <2763@puff.cs.wisc.edu>, brian@cat50.CS.WISC.EDU (Brian Miller) writes:
}>    Regional dialects are rapidly evaporating as mass communication becomes
}> an integral part of modern society. 
}
}This is just not true. It is still impossible to talk about *the*
}British accent, let alone say that British and American accents are
}similar.  Just listen to someone from Glasgow, or Birmingham, or
}Cornwall, or North Wales, or Norwich, or South London or...  And
}that's just in one small island! ....

True, but consider the situation before the advent of mass communication:
That "... one small island ..." used to have over 150 _mutually
incomprehensible_ dialects.  Before radio and the BBC you had to learn a
different language just to visit the next shire.

These days I occasionally have to translate an obscure slang term when my
Cockney friends come to visit (an extreme case).  However, the residents
of the British Isles have little problem understanding each others' spoken
English and the major differences are _accents_ not dialects.  While it's
certainly true that not everyone there speaks "BBC Standard English", it's
equally true that everyone there understands it and speaks so that others
can understand them.

-- 
The Polymath (aka: Jerry Hollombe, hollombe@ttidca.tti.com)  Illegitimati Nil
Citicorp(+)TTI                                                 Carborundum
3100 Ocean Park Blvd.   (213) 452-9191, x2483
Santa Monica, CA  90405 {csun|philabs|psivax}!ttidca!hollombe

rwojcik@bcsaic.UUCP (Rick Wojcik) (05/16/89)

In article <5832@cs.Buffalo.EDU> lammens@sunybcs.UUCP (Jo Lammens) writes:

>In fact there is no such thing as THE B.E. or A.E.  pronunciation, or
>even THE pronunciation for any one dialect. If one acoustically
>analyses people's pronunciation of the same words, the differences
>tend to be very large even among speakers of the same dialect. Yet no
>one really notices. That is one of the reasons that automatic speech
>recognition is so difficult: the identification problem for phonemes
>is a complex one, but you don't realize it until you start measuring
>things or try to build a system that does it. We humans can easily
>shift our perceptual category boundaries around without even being
>conscious about it.

If you make a technical distinction between speech recognition (i.e.
recognition without NLP*) and speech understanding (i.e. recognition with
NLP), then it is a bit easier to understand why speech recognition is not very
feasible across dialect boundaries.  Different dialects have different
phonemic representations for the same morphemes, and you can usually establish
phonemic correspondences only after you have done a lot of higher level
processing.  The problem is that it is very difficult to know when you have a
phonetically motivated distortion of a single underlying sound, or simply a
different phoneme.  Even within the same dialect, phonetic distortion
associated with varying styles and tempoes can obscure the recognition of the
same underlying phonemic string.  It can also cause radically different
phonemic strings to have the similar or identical surface phonetics.  For
example, the word 'cigar' can be pronounced casually in such a way that it is
identical with 'scar'.  This makes it virtually impossible to base speech
recognition on acoustic input alone.

*NLP = natural language processing (syntactic, semantic, pragmatic)
-- 
Rick Wojcik   csnet:  rwojcik@atc.boeing.com	   
              uucp:   uw-beaver!bcsaic!rwojcik 

rwojcik@bcsaic.UUCP (Rick Wojcik) (05/16/89)

Brian Miller (BM) writes:
>>...it is now straightforward to spell/read Russian words.  If you
>>can pronounce a word in Russian, you can spell it.
BM> Any written language that is *not* this regular is a pile of sh`t in my
BM> self-righteous mind.  SHAME on the ineffectual linguists for being to
BM> candy-@ssed to overcome the politics that have prevented such a revamping
BM> as was done in the USSR.

First of all, Russian spelling reform came on the heels of a major political
revolution.  Russian linguists, macho as they were, played a very minor role
in the revolution, which made spelling reform a political possibility.  Even
so, they were unable to get all of their reforms implemented.  The spelling
reform of English, even if it were desirable, is not politically possible.  No
single nation has control over English spelling.  Any attempt to assemble a
conclave of nations to reform spelling would run into a firestorm of
controversy over whose phonemic system should be the basis of the reform.
This would be true to a lesser degree if the reform movement were limited to a
single country.  The likely outcome of an international effort would be a
system that everyone would hate.  I doubt that we Americans would get to
retain syllable-final /r/ in spelling, for example.  So a simple word like
'farm' would have to be spelled something like 'fam,' much to the dismay of
standard American speakers, who scarcely realize that the /r/ is missing in
many British and American dialects.  I can think of no better way to rekindle
an Anglo-American war than to attempt spelling reform.  And nobody is even
talking about the Australians having a say in all this.  Do you really want to
have to confront all those diphthongs, Mayte?  ;-)

BM>    You're more than welcome to be insensitive to the illogic of the English
BM> language.  *I* would jump at the opportunity to reform it.

If you didn't jump, it's very likely that we'd have to toss you overboard. :-)
 
BM>    Regional dialects are rapidly evaporating as mass communication becomes
BM> an integral part of modern society.  I don't give a hoot as to *whose*
BM> pronounciation system we adopt ...so long as its logical and regular.

In that case, let's go for Mark Twain's system.  That was as good a reductio
ad absurdum of spelling reform as I've seen.


[For those interested in Russian spelling only:]

Raymond Shaw (RS) writes:

RS> ...I'm curious; it's been a few years since I studied Russian.
RS> I just can't think of the "etc. etc. etc."  As I remember, it's only "o",
RS> and the rules for pronouncing that vowel are regular:  stressed is one, one
RS> syllable away from stressed is another, and further away from the stressed 
RS> syllable is the third (and only other).  

The word for 'good' is spelled something like 'xorosho' in Russian, with three
'o' vowels.  Stress is on the final /O/ (open /O/ similar to the vowel in
'horse').  It is pronounced [x@rashO], where @ = schwa ('u' in 'luck') and a =
the vowel in 'mom'.  A good way to think of Russian vowel reduction is as
follows:  replace all unstressed /O/ vowels with [@].  Then replace [@] with
[a] if it precedes a stressed syllable or is in absolute initial position.  So
the word for 'cities' is spelled 'goroda' and pronounced [g@rada] with stress
on the final syllable.  The word for 'garden' is spelled 'ogorod' and
pronounced [agarOt] with stress on the final syllable.  Unstressed high vowels
do not reduce in Russian, and reduced vowels after palatalized consonants
undergo fronting.

Russian is not spelled in a totally regular fashion, although it is much more
regular than English.  Besides vowel reduction, there are numerous cases of
'silent' letters--e.g. the word 'hello' is spelled 'zdravstvuyte' but
pronounced variously [zdrastvuyte], [zdrasvuyte], or just [zdraste] (with
palatalization on /t/ not indicated).  The spelling 'Ivanovich' is pronounced
[ivanovich] when used as a last name, but [ivanich] when used as a middle
name.  And so on.  The two vowels /i/ and /y/ (yerih) are really perceived as
the same sound by Russians.  They are allopones in complementary distribution.
True spelling reform would have replaced them with a single symbol, just as
the pre-revolutionary 'jat' was coalesced with the letter 'e'.
  
-- 
Rick Wojcik   csnet:  rwojcik@atc.boeing.com	   
              uucp:   uw-beaver!bcsaic!rwojcik 

jmast@unix.cis.pittsburgh.edu (John M Allen) (05/18/89)

In article <11726@bcsaic.UUCP> rwojcik@bcsaic.UUCP (Rick Wojcik) writes:
>                The two vowels /i/ and /y/ (yerih) are really perceived as
>the same sound by Russians.  They are allopones in complementary distribution.
>True spelling reform would have replaced them with a single symbol, just as
>the pre-revolutionary 'jat' was coalesced with the letter 'e'.

   There is an extremely good reason that the Russian spelling reform
did not combine the /i/ and /y/.  Historically, the two were separate
and the /i/, but not the /y/, caused the palatalization of the
preceding consonant.  There was then a reanalysis so that the
consonant caused the vowel to be front rather than the vowel causing
the consonant to be palatalized.  This reanalaysis means that the
number of consonant phonemes has approximately doubled.

   If the Soviets had merged /i/ and /y/ into one letter, then they
would have had to create about twenty new letters for the palatalized
consonants.  The major consequence of this is that they would have had
to have invested a lot of money into changing the type sets for
printing presses.  It is much easier to remove characters (just don't
use them) than to create new ones.

>Rick Wojcik


\		 |  |			John Allen
 \		\ ||			allen@mercutio.lcl.cmu.edu
/ \					formerly allen@mercurio.lcl.cmu.edu
					jmast@unix.cis.pittsburgh.edu

   "The bond that links your true family is not one of blood, but of
respect and joy in each other's life.  Rarely do members of one family
grow up under the same roof."  -Richard Bach, _Illusions_

rwojcik@bcsaic.UUCP (Rick Wojcik) (05/19/89)

In article <18088@unix.cis.pittsburgh.edu> jmast@unix.cis.pittsburgh.edu (John M Allen) writes:
>   If the Soviets had merged /i/ and /y/ into one letter, then they
>would have had to create about twenty new letters for the palatalized
>consonants.  The major consequence of this is that they would have had
>to have invested a lot of money into changing the type sets for
>printing presses.  It is much easier to remove characters (just don't
>use them) than to create new ones.

That is quite true, and undoubtedly the reason that 'i' and 'y' were not
merged into a single letter.  The fact is that vowel letters in modern Russian
have come to signal phonemic palatalization on preceding consonants.  So,
although the [i]~[y] alternation is completely predictable in speech, it would
not be in writing unless palatalization were indicated on consonant letters.
An interesting fact for the Ripley's Believe It or Not of linguistics.  :-)
Note that true spelling reform of Russian would have created those new
consonants, but the result would have left modern Russians with a severely
impaired ability to read texts printed before the 1920's.  This is just
another of the many reasons why radical spelling reform is not the good idea
that most people think it is.

-- 
Rick Wojcik   csnet:  rwojcik@atc.boeing.com	   
              uucp:   uw-beaver!bcsaic!rwojcik 

jc@minya.UUCP (John Chambers) (05/19/89)

In article <11726@bcsaic.UUCP>, rwojcik@bcsaic.UUCP (Rick Wojcik) writes:
>			...  I doubt that we Americans would get to
> retain syllable-final /r/ in spelling, for example.  So a simple word like
> 'farm' would have to be spelled something like 'fam,' much to the dismay of
> standard American speakers, who scarcely realize that the /r/ is missing in
> many British and American dialects.  

Not true at all.  A reasonable English phonemic spelling system would include 
this /r/ in the spelling, with a note that some dialects (British RP, New 
England, etc.) have a deletion rule for non-pre-vocalic /r/.  This is no more
a problem than, say, the German and Russian rules for final devoicing (which
don't actually apply to quite all the dialects of either language.)  Likewise,
you would spell initial /h/, and give deletion rules.

You also need an r-insertion rule for some English dialects (including both
RP and NE); that need not be indicated in the spelling, because it is quite
predictable from the environment.  Non-pre-vocalic /r/ is not predictable,
so it must be indicated.

The argument that "you can't have a logical spelling system for a language
with dialects" is an old one, and is a total red herring.  Almost all 
languages have dialects; this is no barrier at all to having a logical
spelling system.  The basic approach is to spell things that are in any
of the major dialects, and have a list of pronunciation rules.

> [For those interested in Russian spelling only:]
> 
> The word for 'good' is spelled something like 'xorosho' in Russian, with three
> 'o' vowels.  Stress is on the final /O/ (open /O/ similar to the vowel in
> 'horse').  It is pronounced [x@rashO], where @ = schwa ('u' in 'luck') and a =
> the vowel in 'mom'.  A good way to think of Russian vowel reduction is as
> follows:  replace all unstressed /O/ vowels with [@].  Then replace [@] with
> [a] if it precedes a stressed syllable or is in absolute initial position.  So
> the word for 'cities' is spelled 'goroda' and pronounced [g@rada] with stress
> on the final syllable.  The word for 'garden' is spelled 'ogorod' and
> pronounced [agarOt] with stress on the final syllable.  Unstressed high vowels
> do not reduce in Russian, and reduced vowels after palatalized consonants
> undergo fronting.
 
Actually, this isn't quite correct; it gives the wrong value for written 'a'
in noninitial, nonstressed position.  You need a third rule, or use a better 
formulation:
	1. Rewrite unstressed 'o' as 'a'.
	2. Pronounce unstressed 'a' as /@/ if not initial or prestress.
I've seen both formulations in texts; the latter is simpler.

There's also a separate set of rules when 'o', 'a', or 'e' follow a soft
consonant, of course.

-- 
--
All opinions Copyright 1989 by John Chambers; for licensing information contact:
	John Chambers <{adelie,ima,mit-eddie}!minya!{jc,root}> (617/484-6393)

goer@sophist.uucp (Richard Goerwitz) (05/21/89)

In article <5@minya.UUCP> jc@minya.UUCP (John Chambers) writes:
>A reasonable English phonemic spelling system would include 
>this /r/ in the spelling, with a note that some dialects (British RP, New 
>England, etc.) have a deletion rule for non-pre-vocalic /r/.

The assumption here seems to be, e.g.:

     /bar/ [ba:]

because the /r/ is realized as [r] in forms like "barring."

As a language person, rather than a true theoretical linguist, I must
confess to some annoyance at the pervasiveness of this patently Pla-
tonic view of human language.  How do we know that "bar" and "barring"
represent divergent hypostases of a single underlying form (in the
Platonic sense)?  Why not have speakers simply lexicalize the relation-
ship?  Even if speakers relate the two in a more rule-oriented manner
than this, why not simply see the two in a more Aristotelian light -
as *related* rather than as manifesting some underlying, abstract form?
I see no logical connection between rules relating two forms and the
existence of an abstract *common* form.  I also do not see why both
models cannot coexist.

I would appreciate being enlightened on this point by those who are more
interested in general linguistic theory.  It is something that has bugged me
ever since my introduction to linguistics as an undergraduate in that
field at the University of Chicago.

                                       -Richard L. Goerwitz
                                       goer@sophist.uchicago.edu
                                       rutgers!oddjob!gide!sophist!goer

jmast@unix.cis.pittsburgh.edu (John M Allen) (05/23/89)

In article <3368@tank.uchicago.edu> goer@sophist.UUCP (Richard Goerwitz)
writes:
>The assumption here seems to be, e.g.:
>
>     /bar/ [ba:]
>
>because the /r/ is realized as [r] in forms like "barring."
>
[...]
>                               How do we know that "bar" and "barring"
>represent divergent hypostases of a single underlying form (in the
>Platonic sense)?  Why not have speakers simply lexicalize the relation-
>ship?

   There are several reasons that a single common base form is
prefered over several forms related by rules, but the most important
one is memory size.  Humans have a large but finite amount of memory.
If you assume that humans store a single lexical form for regualar
words and all of the spoken forms are derived by a small number of
rules then you have a fairly small memory load.  On the other hand if
you store all possible forms that a word might have the memory load is
dramatically increased because each morpheme may have several
different morphophonemic shapes.

   Consider the following data from Russian verbs.  According to one
analysis, there are several rules that operate on verbs.
	1)  vocalization.  A nasal between two consonants will change
to "a".
		z^m + t' ==> z^at'
		z^m + u  ==> z^mu

	2)  sonorant deletion.  A sonorant, [n,m,j,v], at the end of a
root is deleted when the suffix begins with a consonant.
		znaj + t' ==> znat'
		znaj + u  ==> znaju

	3)  palatalization.  Some of the verbal endings palatalize the
preceding consonant.
		z^m + u  ==> z^mu
		z^m + om ==> z^m'om


   Now comparing the two hypotheses.

	      Base+Rules		    Lexicalization
			Cumulative			Cumulative
	Forms	  	  Total		Forms		  Total
---------------------------------------------------------------------
Rules			    3				    0

z^m	z^m		    4		z^a,z^m,z^m'	    3
  `press'
znaj	znaj		    5		zna,znaj	    5
  `know'
nac^n	nac^n		    6		nac^a,nac^n,nac^n'  8
  `learn'
z^iv	z^iv		    7		z^i,z^iv,z^iv'	   11
  `live'

As you can see with only four words, the lexical approach takes more
space than derivations from a base.  Furthermore, every Russian verb
has as least two different morphophonemic shape under this analysis,
so the difference will continue to increase as you add more verbs to
the lexicon.

   Another reason that the single form is preferred is that many
alternations are automatic.  If you give a person a new word in some
form and ask for the same word in another form, they will apply all
the appropriate rules.  This is very difficult to explain using a
lexical approach because the person could not have heard the form
before and thus could not have stored it.  The rule based approach has
no problem explaining this.

   Finally, lexicalized items tend to resist change.  If every word
had all of its forms lexicalized, then we would expect many more
exceptions to the morphophonemic rules than there are.  The rule based
approach simply changes the rule in the appropriate way.

> -Richard L. Goerwitz

\		 |  |			John Allen
 \		\ ||			allen@mercutio.lcl.cmu.edu
/ \					formerly allen@mercurio.lcl.cmu.edu
					jmast@unix.cis.pittsburgh.edu

   "The bond that links your true family is not one of blood, but of
respect and joy in each other's life.  Rarely do members of one family
grow up under the same roof."  -Richard Bach, _Illusions_

goer@sophist.uucp (Richard Goerwitz) (05/24/89)

John M. Allen writes:
>>                               How do we know that "bar" and "barring"
>>represent divergent hypostases of a single underlying form (in the
>>Platonic sense)?  Why not have speakers simply lexicalize the relation-
>>ship?
>
>   There are several reasons that a single common base form is
>prefered over several forms related by rules, but the most important
>one is memory size....
>[If] you store all possible forms that a word might have the memory load is
>dramatically increased because each morpheme may have several
>different morphophonemic shapes.

Same old stuff I've been hearing.  You assume that Ockham's razor applies
in a straightforward way to the very elusive processes of the human mind.
Sure, it works out nice for your theories, but does it in fact reflect the
underlying psychological reality?

>   Consider the following data from Russian verbs... (examples deleted).
>
>As you can see with only four words, the lexical approach takes more
>space than derivations from a base.  Furthermore, every Russian verb
>has as least two different morphophonemic shape under this analysis,
>so the difference will continue to increase as you add more verbs to
>the lexicon....

>   Finally, lexicalized items tend to resist change.  If every word
>had all of its forms lexicalized, then we would expect many more
>exceptions to the morphophonemic rules than there are.  The rule based
>approach simply changes the rule in the appropriate way.

Let's get into the question of how one might relate sets of forms.  Say
we have the word (taken from biblical Hebrew, * = attested in older
dialects preserved in Greek, Latin, or Akkadian transcription; the verb
itself does not occur in this precise form - it is a junk form Semitists
use to elucidate the CV pattern characteristic of the so-called Dt
stem):

         *hithpa``alu: (`=voiced pharyng. fric.; th = voiced interdent. fric.)

Now say that we have two forms of this verb.  One is clause-final (a)
and one is clause medial (b).  The tonic syllable in the clause-final
form, if open and short, is lengthened (i.e. the feature +short is neu-
tralized in open, "pausally" stressed syllables):

         a) hithpa``a:lu:  (paroxytone)
         b) *hithpa``alu:  (paroxytone)

Both of these forms are accented on the penult.

Now what if we have an accent shift in forms with a final long vowel
(V:# whose penult is open and short):

         a) hithpa``a:lu: (still paroxytone)
         b) hithpa``alu:  (now oxytone)

And then we get our pretonic a in form (a) going to shewa, and our
tonic /a:/ in (a) going to a low back rounded vowel (the language
is shifting to stress timing, and is losing length distinctions).
Labialization and development of a shewa are natural offshoots thereof.
The result is a pausal form (a) that differs dramatically from its
nonpausal (i.e. clause-medial) counterpart:

         a) hithpa``clu  (c = low back rounded vowel)
         b) hithpa``@lu  (@ = shewa)

I'm not kidding, by the way, this all really happened.  As I mentioned
above, we have Greek, Latin, Akkadian transcriptions, as well as sup-
porting evidence from cognate dialects.  The last two forms are what
is attested in biblical Hebrew.

Generative treatments of Hebrew analyze these two forms, (a) and
(b), as underlyingly the same.  To get from an abstract form to
its realization is very, very complicated.  I would refer you, for
instance, to several MIT theses on this subject (Prince, Rappaport,
McCarthy). 

Now here's the catch:  Words fall into patterns.  So if we simply
have a rule that the hithpa``@lu pattern goes to hithpa``clu in
pause, we cover many, many similar verbs in Hebrew.  In fact, it
is much simpler to lexicalize the PATTERNS here than to bother re-
capitulating synchronically the diachronic changes 1) pausal lengthening,
2) accent shift, 3) labialization of /a:/, 4) a -> @ / C_[-stress]CV,
which would be necessary if we were to adopt the theory of an under-
lyingly common, Platonic "form."  Ultimately, it is easier for the
speaker to simply follow a simple conversion rule than to work off of
a common form.

As for the psychological reality of the conversion rule (and the im-
plied lack of a common form), I would point to certain wonderfully
revealing cases of analogy.  The verb shown above has the singular

       c) hithpa``el

Recall that the form given above went *hithpa``alu:.  This is actually
a plural form, with the masculine plural morpheme -u tacked onto the
end.  The preceding vowel is an /a/.  This is in fact what we would ex-]
pect in the singular -

       c) *hithpa``al  (attested in older transcriptions)

Clearly some sort of shift has occurred.  However, it has no parallels
elsewhere in the language.  In fact, the expected trend is towards low-
ering of stressed, high vowels.  What is going on here?

The answer is that the so-called D-stem has become the model on which
a new Dt-stem has begun to be pronounced:

       d) pa``el (D-stem)
       c) hithpa``el (new Dt-stem [originally a passive/reflexive of the D]) 

The interesting thing here is that the pausal form retains its historical
vowel:

       e) hithpa``cl (c = low back rounded vowel < /a:/ < stressed /a/)

Why, if we have a single, underlying form, do we have re-patterning with the
context form, and not with its pausal counterpart.  The pa``el pattern has
a pausal form of its own (which in this instance looks like the context form,
but sometimes differs).

What I am getting at is that in Hebrew, it appears that native speakers had
no concept of an abstraction unifying context forms with their pausal counter-
parts.  Hence you could have analogy with the one without involving the other.

In broader terms, one might observe that Semitic languages generally are
very patterned, in the sense of having a fairly limited and predictable
set of nominal and verbal patterns.  In these languages it is much simpler
to derive corresponding plural/singular, passive/active, pausal/context
etc. forms by simple pattern-replacement rules.  To derive them from under-
lying common forms is much more complicated, especially in cases where the
language in question has become stress-timed (Hebrew, Aramaic). 

The point I personally have gleaned from all of this is that the generative
model is not a universal theory (big suprise).  It may be valid for some
languages, and for some parts of other languages.  It may also coexist with
other strategies.  No one ever said that people were consistent (except may-
be economists and linguists - both guilty of practicing "pseudo-science").
No one ever said that Plato (or as some would have it, Descartes) was right.
But then no one ever said Aristotle was right, either.

From my limited standpoint, it appears that generative grammar is really
just another incarnation of the "God's truth" school of linguistics - one
theoretical extreme in an infinitely varied and probably ultimately indes-
cribable spectrum of strategies we human beings might possible use to com-
municate verbally.

                                       -Richard L. Goerwitz
                                       goer@sophist.uchicago.edu
                                       rutgers!oddjob!gide!sophist!goer