minimal pairs are not always there

Robert Whiting whiting at cc.helsinki.fi
Sun Oct 15 11:46:47 UTC 2000


On Tue, 25 Apr 2000 Dr. John E. McLaughlin <mclasutt at brigham.net> wrote:

<snip>

>[Robert Whiting wrote]
<snip>
>> I think that you truly have to consider both 'teeth' and 'teethe'
>> as morphophonemic variants of 'tooth'.

>This is not a productive rule of Modern English.  You really must
>distinguish between diachrony and synchrony.  I can no longer
>make a noun into a verb by adding Germanic *-jan to it and then
>change the vowel by umlaut.  It hasn't been a productive rule for
>about a thousand years.

And you really must distinguish between productive rules and
other rules.  "The" productive rule is the one that is used to
create new forms in the language.  It is the default rule when a
form is not marked for some other rule.  Other rules are
exceptions to the productive rule.  These are markedness rules
because forms that follow them have to be marked in the lexicon.
Speakers of a language learn the markedness rules when they learn
the language.  There must be synchronic rules that produce these
forms or else they wouldn't exist.  The synchronic rules aren't
the same as the diachronic rules that originally produced the
forms, but they have the same result.  They have to, because the
basis of both historical grammar and synchronic grammar is the same:
the language as it is presently spoken.  But modern speakers don't
know that this is an umlaut rule (most modern speakers don't even
know what umlaut is); they just know it as a rule that six words
in the lexicon are marked for.  If they didn't know the rule, the
plural of 'foot' would be 'foots' and the plural of 'tooth' would
be 'tooths'.

Linguists tend to use "productive" in two senses with respect to
rules.  "The" productive rule is the default rule that is used to
produce new forms in the language.  "A" productive rule is one
that the speakers of the language know and use to generate some
forms.  So a rule may not be "the" productive rule but may still
be "a" productive rule in the sense that speakers use it to
produce forms currently used in the language.  A non-productive
rule is one that is no longer used at all in the language.  An
example of a non-productive rule in English is the use of
'-(e)st' to form second person singular verb forms.  There are no
second person verb forms in '-(e)st' (like 'goest' or 'didst') in
modern English because this rule is now non-productive.  There
are, however, noun/verb pairs like grass/graze, house/house,
half/halve/, breath/breathe, bath/bathe, etc.  This indicates
that there is a productive morphophonemic rule that generates
these forms.  If there were no productive rule that creates these
forms, then the forms wouldn't exist synchronically.  The verb
would be the same as the noun following "the" productive rule.

If you are trying to tell me that rules that aren't "the"
productive rule aren't known by native speakers then you are
trying to tell me that there aren't any strong verbs in English
because the rules that produced them are no longer productive, I
just don't believe it.  You may go around saying things like "I
seed it with my own eyes" or "I have always thinked so," but most
of the native speakers I know do not.  If you think that
synchronic grammar consists of only "the" productive rules, then
you have a very idiosyncratic view of synchronic grammar.
Synchronic grammar consists of both "the" productive (default)
rules and other productive (markedness) rules.  The markedness
rules generate exceptions to "the" productive rules.

>Therefore they are NOT synchronic morphophonemic variants.

Sure they are.  Any time you have a systematic phonological
change that results in a systematic morphological change
you have morphophonemic variation.  The fact that modern
speakers don't know that it is a morphophonemic change doesn't
alter the linguistic fact.  Again, most modern speakers don't
know what morphophonemics is.

There are two basic types of morphophonemic alternation.  One
occurs when a single morpheme has different phonological shapes
depending on its environment.  An example is the English plural
morpheme '-(e)s' which is either voiced or voiceless depending on
its phonological environment.  Another example is the variation
of the determiner 'a/an' depending on its phonetic environement.
The other type of morphophonemic variation occurs when a regular
phonological change causes a regular morphological change.  Such
changes often result in a change in grammatical meaning but not
in lexical meaning or move the word from one functional category
to another without significantly changing its underlying meaning.

You have this exactly backwards.  They *are* synchronic
morphophonemic variants.  They are not archaic morphophonemic
variants.  The process that created them was not morphophonemic
in nature.  It was a simple phonological change.  It was
triggered by a phonological environment that had no regard for
meaning.  But the synchronic rule that maintains these forms in
the language is morphophonemic because the environment that
triggers these changes is now morphological rather than
phonological.  Phonemes are units of sound.  Morphemes are units
of meaning.  When a specific and predictable change in a sound
causes a specific and predictable change in meaning, or when a
morpheme has more than one phonetic realization on a predictable
basis, then you have morphophonemics.

>[I wrote]

>>> Historically, yes, these two forms were not (the 'e' on the end
>>> of teethe was a phonetic element which put the voiceless /th/ in
>>> a voicing environment, but synchronically, there is no
>>> distinction between the two except for the final voicing of th/dh
>>> (the lengthening of [i] in 'teethe' is due to the voicing of dh,
>>> it does not cause the voicing).

>[Robert wrote]

>> Historically, this is nonsense.  the lengthing of [i:] in
>> 'teethe' is a matter of stress.  It is a matter of vowel
>> quantity, not vowel quality.  Both 'teeth' and 'teethe' have [i:]
>> and if the ending is not stressed, both have the same vowel
>> quality.  The [i:] in both 'teeth' and 'teethe' is the result of
>> umlaut caused by the addition of the plural ending (beginning
>> with '-i') and the verbal suffix (beginning with '-j'; exactly
>> the same change that took place in 'doom' - 'deem'),
>> respectively.  The fact that many speakers introduce this
>> additional distinction by stressing the ending of the verb
>> suggests that they do not consider the [th] - [dh] distinction to
>> be sufficient (i.e., they do not consider it phonemic).  If
>> 'tooth' had not preserved its umlaut plural (i.e., if 'tooth' [+
>> plural]  --> *'tooths'), the question wouldn't arise.

>Sorry, Robert, but you're mixing up all kinds of
>diachronic/synchronic and phonetic/phonemic levels here.  Ask any
>phonetician of English and he or she will gladly tell you that
>vowels in Modern English preceding a voiced consonant are
>measurably longer in duration that vowels preceding a voiceless
>consonant.

Yes, this last statement is true (even if the first one isn't :>).
It is a matter of timing.  Since vowels are voiced, making a
voiceless consonant sound following a vowel involves turning
voicing off *before* beginning the consonant sound, creating a
clear boundary between the vowel and the unvoiced consonant.
Voiced sounds do not turn off voicing so there is no clear
boundary between the vowel and the consonant.  As a result the
vowel sounds longer (and can be measured to be longer) because it
continues into the beginning of the consonant.  But I wouldn't
call this lengthening (at least, not with respect to English), I
would call it prolongation (a trivial-seeming point, but it
avoids misunderstandings).

>I'm not at all talking about the "long/short" vowel distinctions
>of Old English, nor the diachronic processes that you believe are
>still operating in Modern English morphology and morphophonemics.

No, I can see that you aren't now; it was your use of
"lengthening" that misled me.  But you are the one who seems to
be having problems separating out the diachronic and synchronic
processes that produce these forms.  I do not believe that the
processes that originally produced these forms are still
operating in modern English.  The original creation of these
forms was not morphophonemic.  It was a simple sound change.

English once had no voiced spirants.  It had only unvoiced /th/,
/f/, and /s/ (Proto-Germanic *z merged with *r in pre-English).
Then a phonetic change took place that can be expressed by the
rule [unvoiced spirant] > [voiced spirant] / [voice]___[voice].
There are other factors that may affect the change, such as the
location of the stress, but note that this change is entirely
phonological:  the conditioning environment is phonetic and it
affects all sounds in that environment regardless of meaning.
This change caused the voicing of spirants in plural forms and
verbal infinitives, not because of meaning, but because these
forms, as they were at the time of the change, provided the
required environment for the change to take place.  This change
did not create any new contrasts (phonemes).  All it did was
create voiced allophones of the original phonemes: /th/ [th, dh],
/f/ [f, v], /s/ [s, z].

The forms of the plural and the infinitive also provided
environments for a number of other changes (rules) to take place
(although not at the same time) such as partial assimlation of
vowels (umlaut) and lengthening of short vowels in open
syllables.  All in all, quite a number of rules pile up on these
forms.  Then other rules cause the forms to change, eliminating
the conditioning environment that brought about these changes.
However, the unvoiced/voiced distinction in spirants was kept in
plurals (now formed with '-s') and noun/verb pairs.  The new rule
looks like this: [final unvoiced spirant] --> [voiced spirant] /
[PLURAL, VERB].  Note that the phonetic environment has been
replaced with a grammatical one.  The change is no longer
predictable phonologically.  But it is still predictable.  It is
only when the change is used outside this enviroment that it is
no longer predictable (in view of one of your previous posts,
note that [POSSESSIVE] is not part of the environment; thus the
distinction of paths/path's is not phonemic [phonetic, yes;
phonemic, no] since the [dhz] of 'paths' is predictable by rule
as is the [ths] of 'path's').

But the modern synchronic rule (as stated above) is
morphophonemic and, of course, it is not "the" productive rule.
The productive rule is NOUN --> VERB / ... (meaning that any noun
can be used as a verb).

But if you believe that the modern language doesn't have a rule
that makes 'teethe' the verb from 'tooth', just start asking
people if their baby is toothing and see what kind of reactions
you get.  Or if you believe that there is no modern rule that
makes 'thieves' and 'halves' the plural of 'thief' and 'half',
then you can use plurals like 'thiefs' and 'halfs'.  People will
probably understand you, because you are using "the" productive
rule, but they will also probably wonder what your native
language is.

>Phonemically, both teeth and teethe have /i:/ (or /i/ if you
>prefer distinguishing between /i/ and /I/ rather than using
>length as the distinguishing feature), but phonetically, the /i:/
>of 'teeth' is not as long as the /i:/ in 'teethe'.  There's no
>debate about this among phoneticians.

Yes I agree that the /i:/ of 'teethe' can be expected to be more
protracted than that of 'teeth' because of the voicing of the
final consonant.  But in my view, it is protracted even beyond
what one would expect based on similar pairs (like 'safe', 'save'
or 'staple', 'stable') leading me to believe that there is
additional stress on the form.  One notes a similarity in words
such as 'writer' /rait at r/ and 'rider' /rai:d at r/ (where : implies
additional vowel quantity).  In dialects where /t/ and /d/ have
merged into a dental flap, this additional vowel quantity is now
the phonemic difference (/raiFLAP at r/, /rai:FLAP at r/).  Since this
kind of distinction is not normally phonemic in English, there may
have been a rephonemicization ([shortness]+FLAP ~ /*t/ and
[longness]+FLAP ~ /*d/).

[I wrote]

>> But there are several good minimal pairs in (at least American)
>> English for th/dh--ether/either, thigh/thy, wreath/wreathe,
>> sooth/soothe, etc.

>[Robert wrote]

>> 'ether' [borrowed word] - 'either' [native word]
>> 'thigh' [non-pronoun] - 'thy' [pronoun]
>> 'wreath' [noun] - 'wreathe' [verb]
>> 'sooth' [noun] - 'soothe' [verb]

>Both here (and in your previous posts) you are marking way too
>many words in English as "borrowed".

Golly, one word is "way too many." :)

But here is the tricky bit about 'ether':  This word was already
in Middle English (first attested in the 14th century according
to my dictionary).  This is before any possiblity of a phoneme
split between [th] and [dh].  At this stage, [th] and [dh] were
clearly still allophones because the original conditioning
environment that caused the sound change had not yet fully
disappeared.  In this situation, 'ether', with its intervocalic
[th] would have stuck out a mile as a foreign word.  And 'either'
didn't get its present phonological shape until after the Great
Vowel Shift, so there was no need to distinguish 'ether' from
'either' originally.  Even after the Great Vowel Shift, [th] and
[dh] were still not phonemically distinct so there is no way to
claim that 'ether' and 'either' were originally differentiated by
different phonemes.

So you are trying to tell me that when [th] and [dh] allegedly
split into two phonemes, the difference between 'ether' and
'either' suddenly became a phonemic distinction.  My question
then is, how were they distinguished during the centuries before
this split took place?  My answer would be that /th/ had two
allophones, [th] and [dh]; intervocalically the rule was that
/th/ was realized as [dh] in native words and as [th] in
non-native words that had [th] in this position.  Since this rule
still operates today, it is hard to see how the sole example
where its operation creates a minimal distinction can be
considered as evidence of the phonemicity of the segements
involved.

>There are many words in English that are clearly marked as
>"borrowed" in the usage of most speakers (you mentioned 'padre',
>for example), but you are not at all careful in drawing the line
>between words that are perceived and used as borrowed terms and
>words that have been completely Anglicized.  Should we mark
>'copper', 'mint', 'mile' and 'church' as "borrowed"?  Or how
>about 'seal (the animal)', 'auk', 'herring', and 'sea'?

You are quite correct that "nativization" is in the perception of
the speakers.  But all the words you list were already present in
Old English (and I'm not so sure about 'seal' [the animal],
'herring', and 'sea' being loans; you'd do better with 'egg',
'cup', and 'rose'), and as you say, have been completely
Anglicized.  And there is not an intervocalic [th] in the lot.

To check on perceptions of "borrowed" vs. "native", let's look at
three borrowings of the same root from the Latin-French
continuum.  If we consider 'cant', 'chant', and 'chanson' we will
see that these three words came into English at various stages of
Latin-French phonological development as indicated by the
pronunciation of the initial consonant.  'Cant' and 'chant' are
fully Anglicized and probably would not be recognized as loans by
naive native speakers.  But 'chanson' has been in the language
for about 400 years and it still has its French pronunciation.
It simply resists Anglicization because it is not known to most
naive native speakers.  People who know this word are likely to
know that it has a French pronunciation and to know why.

>After a thousand years, borrowed words will have suffered one of
>two fates generally:  1) they will be so few in number that they
>will have been completely adapted to the borrowing language's
>phonology so that they are no longer identifiable as borrowed
>words, or 2) they will be so many in number that they will have
>changed the phonological structure of the borrowing language and
>might be identifiable to a linguist as an old borrowing, but to
>no native speaker.  The latter is the case in English with much
>of our borrowed vocabulary.

This latter process is what Hock, _Principles of Historical
Linguistics_ (1986), 396, calls "nativization by adoption" which
he characterizes as "downright adoption of the foreign segment or
... adoption of the segment in a context in which it does not
occur natively."  Let me quote what he has to say in summary on
pp. 396-97:

   The importance of the Old Irish example is twofold.  First it
   suggests that 'nativization by adoption' is qualitatively
   different from the other nativization processes:  It does not
   really nativize at all, but merely 'admits' foreign words into
   the language without losing track of the fact that they are
   and remain foreign.  It is only after these words have been
   around for quite some time, have been used often enough, and
   in enough different and novel contexts, that speakers may
   slowly begin to lose the feeling that they are not native.
   Nativization by adoption, then, is not an immediate process,
   but one of slow, gradual, even grudging acceptance.

Hock goes on to point out that the second thing that the Old
Irish example shows is that once a word is accepted as native,
then it is treated exactly like a native word.  Now while a
thousand years is surely enough time for nativization to take
place, most of the words that I am talking about have been in
English far less than that.  Surviving words that were already in
OE are surely considered native.  Even words like 'sky' and
'skirt' are doubtless considered native despite the fact that the
initial [sk] tells a linguist that they are loans from Old Norse
and despite the fact that 'skirt' has a native doublet, 'shirt'.

>Take, for example, -(o)logy.  It follows a Greek stress pattern.
>Originally, it was borrowed and only used with Greek stems. It
>soon was also used with Latin stems, but we can now say that
>-ology is completely part of the English "native" vocabulary
>because it is productive with any kind of stem--whether of Greek,
>Latin, English, or Hindustani origin.  One needs only to listen
>to college students talk for any length of time to hear myriads
>of -ology words.  It's a productive native suffix now.

The same is of course true of '-ize' and '-ism', but despite the
fact that they are fully productive suffixes in English, most
speakers probably still perceive them as foreign.  In its
borrowings, English is omnivorous.  It will use foreign suffixes
with native stems and foreign stems with native suffixes, perhaps
knowing that they are foreign, but not really caring.

All in all, the likelihood that borrowed words will be considered
native depends on a number of factors, which include (but are not
necessarily limited to):

1) The length of time that the word has been in the language:
   words that were borrowed already in OE like 'cup', 'rose',
   'street' are likely to be considered native because they have
   undergone most of the same phonetic changes that native words
   have.

2) How well the words resemble native words:  words like 'candle'
   or 'castle' are likely to be considered native because of
   their resemblance to native words like 'handle' or 'wrestle'.
   On the other hand, words like 'chandelier' or 'château', which
   were borrowed from the same roots but at a much later date,
   are almost certain to be considered foreign because of their
   un-English phonology and morphology.  Likewise, 'chief' is
   likely to be considered native because of its resemblance to
   native 'thief', while 'chef', borrowed from the same root but
   about 500 years later will probably be considered foreign.
   Similarly, 'file' (a collection of papers) is likely to be
   considered native because it is homonymous with the native
   'file' (the tool). No connection is likely to be seen between
   the borrowed 'file' and 'filament', which may or may not be
   considered a loan.

   'Arms' (weapons) and 'army' are likely to be considered native
   because of their similarity to the native 'arm', while the
   native word for "army," 'here', has disappeared from the
   language with only a few traces.  Indeed, were the good
   (Old) English word 'heretoga', "war leader," to be revived,
   most people would doubtless consider it foreign (unless they
   recognize it through the German cognate 'Herzog').

3) Short words (particularly those that match English
   phonotactics) are more likely to be considered native than
   long words.  Thus 'air' or 'sky' is likely to be considered
   native, while 'atmosphere' would doubtless be considered
   foreign.

4) In general, words borrowed from languages that are closely
   cognate with English (such as other Germanic languages) are
   more likely to be considered native than words borrowed from
   more distantly related or unrelated languages.  Thus 'egg',
   'fir', and 'skin' are likely to be considered native while
   'khaki', 'pajamas/pyjamas', and 'etiolate' are likely to be
   considered foreign.

5) Common, frequently used words are more likely to be considered
   native than rare words (however valid or false this perception
   may be).

6) Words that flat out violate English phonotactics in a fairly
   spectacular fashion, like 'aardvark', 'zygote', 'syzygy', or
   'quincunx' are likely to always be considered foreign (or at
   least non-native) no matter how long they stay in the
   language.  Loans with intervocalic [th] may very well belong
   to this category (although this is not a particularly
   spectacular violation).

>To do a phonological analysis of a language based on
>establishing an artificial distinction between ancient
>borrowings and so-called "native" words is weak, at best.

It's hardly an artificial distinction.  The distinction is
disgustingly regular.  And the vast majority of these words are
not "ancient" borrowings but occurred after the change took
place.  Words that were in the language before the change took
place (like 'rose') underwent the change.

>If your only criteria for linking [th] and [dh] as allophones of
>a single phoneme in MODERN English is a morphophonemic rule that
>hasn't been productive for over a thousand years,

For the (next to) last time, John, the change that created these
forms was not morphophonemic.  There is a productive modern
English rule that creates these forms for modern speakers or else
the forms wouldn't be there.  The synchronic rule is not the same
one that created the forms.

>and a distinction between very old borrowed words and "native"
>words, then you haven't proven the relationship.  There is,
>indeed, a diachronic relationship between the two, and the two
>were, indeed, allophones of a single phoneme in Old English.  But
>in Modern English, the two have split into two phonemes.

Sure they have split, possibly even into two phonemes.  There is
no reason why [th] and [dh] shouldn't or couldn't be separate
phonemes in English.  They just aren't used as separate phonemes;
that is as markers of *arbitrary* distinctions in meaning.  Show
me an example where the difference can't be accounted for by a
fully generalized rule and I'll believe it.

But you are also forgetting about the evidence of loan words from
languages that do have a /dh/ phoneme where this phoneme doesn't
get taken into English as [dh] (e.g., 'dhow' [from Arabic] or
'dharma' [from Hindi]).  If [dh] is really a phoneme in English,
then there is no reason why these words shouldn't come into
English with that phoneme.  After all, foreign words with [v] or
[z] come into English with /v/ or /z/.  So why shouldn't foreign
words with [dh] come into English with /*dh/?

>[I wrote]

>>> However, because of the very complex morphophonemics of Central Numic
>>> and the historical changes that have further obscured them in
>>> Comanche, this language is full of pairs that look very much like
>>> minimal pairs on the surface, but are not.  For example, [papi] 'head'
>>> and [pavi] 'older brother' look very much like a minimal pair.
>>> However, they represent /pa=pi/ and /papi/ respectively.  (The = is a
>>> phoneme in Comanche that prevents the lenition of a following stop.
>>> It is fully justified on morphophonemic grounds without relying on the
>>> historical presence of /n/ in Panamint and Shoshoni which is cognate.)
>>> There are a bundle of these:  [ata] 'different' /a=ta/ versus [ara]
>>> 'uncle' /ata/, etc.

>[Robert wrote]

>> Fascinating.  Please, sir, what is the phonetic realization of this
>> phoneme [=]?  Oh, I just realized -- it can't have a phonetic
>> realization or else [papi] and [pavi] wouldn't seem to be a minimal
>> pair.  It just blocks some normal phonetic change.  I'm sorry, John,
>> but this looks like a device to create a phonetic environment to
>> explain why some stops don't undergo lenition when the conditioning
>> environment that prevented it has been lost historically.  I'll tell
>> you what:  Let's assume that English has a phoneme (let's call it [=]
>> just for consistency) that prevents an intervocalic dental spirant
>> from being voiced.  Now let's insert this phoneme in a word like
>> 'ether' which shows an unvoiced intervocalic dental spirant /i:=ther/.
>> Good -- now we no longer have a minimal pair 'ether' - 'either'.  Now
>> let's assume that English inserts this phoneme in all loanwords that
>> have an unvoiced dental spirant in a voiced environment.  Voila -- a
>> phonetic environment that explains why loanwords have unvoiced
>> intervocalic [th].  Now all we need is a rule that says /=th/ -->
>> [dh] /__ m# and all intervocalic [th] in English is accounted for by
>> phonological rules.  Hey, this is fun.

>Well, Robert, you've fallen into the trap that countless other
>non-Numicists have blundered into.

Perhaps, but it looks like the trap was dug by the Numicists
themselves rather than being there in the language.

>But it is also illustrative of how different your morphophonemic
>evidence for lumping [th]/[dh] in English is from the Comanche
>problem at hand.  Here's some very basic data to show that /=/
>has a phonemic status.

<snip of lots of good stuff about Comanche morphophonemics>

>By now you should realize that this is not some feature of the
>second element, but a feature of the stem that causes the initial
>consonant of the second element to be preaspirated, nonlenited,
>and lenited.  Unlike the voicing of /th/ to [dh], it is fully
>productive in (at least preobsolescent) Comanche.  There is
>something following each of these nominal stems which is
>neutralized in word final position.  From Shoshoni evidence, we
>know that these "final features" are -C (an undifferentiated
>consonant that causes gemination in Shoshoni and preaspiration
>in Comanche), -n or -= (prenasalization in Shoshoni, a nonlenited
>stop in Comanche), and zero (allows lenition in both Shoshoni and
>Comanche).

There is a similar phenomenon in Finnish, known as gemination
(sometimes referred to as the "phantom consonant").  Certain
forms of certain types of verbs and certain forms of certain
types of nouns trigger gemination of the initial consonant of the
following word (e.g., <tule tänne> "come here" is realized as
'tulet tänne').  This event is both phonologically (always comes
after [e]) and morphologically (only certain forms that meet the
phonological criterion trigger it) conditioned.  It usually
doesn't get a very detailed or coherent discussion in grammars,
but I have never seen anyone try to explain it by a stem-final
silent phoneme.  Perhaps Ante Aikio can help me out on this one.

>Now this does bring up an important point that I'm sure you'll
>agree with. There is not a clear boundary line that demarcates
>when a phoneme has split or when morphophonemic distinctions have
>ceased productivity or when any number of changes have finally
>and irreversibly taken place.  Comanche is a very clear
>borderline case.  The phonemic status of = (Shoshoni /n/) is not
>completely black or white.  Such is also the case with the
>phonemic split between [th] and [dh].

Yes, I agree.

>(Now's the part where we disagree.)

Again, I agree.

>Because of the fully productive nature of the (morpho)phonemic
>final features in Comanche (including /=/), they must be set up
>as phonemes in the language, although admitting that their life
>expectancy is low.

I can't see why.  These are simply morphophonemic alternations.
Such alternations with different stem types are normally treated
as variant declensions based on noun classes.  If you want to
call these variants morphophonemes, but not phonemes, I would
have no objection.

>Because of the completely non-productive nature of the old
>morphophonemic processes which gave rise to [th] versus [dh],

For the last time, John, the processes that gave rise to [th] and
[dh] were not morphophonemic.  It is the modern synchronic rules
that are morphophonemic.  And they are productive because if they
weren't, the forms wouldn't be in the language.

>because [th] in [dh]'s environments has become firmly fixed by
>old loan words that have become nativized, and because the
>voicing environments for [dh] have been lost without the
>subsequent devoicing of [dh] to [th], then we must set up two
>phonemes in Modern English--/th/ and /dh/, although admitting
>that they are only recently distinguishable from one another.

We can set them up, but we can't get speakers to use them.  Until
speakers use them as phonemes we can't say that they exist as
phonemes.  The potential for their use as phonemes exists, and
they are probably on the verge of being used as phonemes (for
example, 'bath' has both forms of the plural [baths] and [badhz];
if these two forms come to have distinct meanings {such as
[badhz] denoting an immersion in water ('the children have all
had their baths[dhz]') and [baths] referring to a place where one
can be immersed in water ('we visited the baths[ths] of Caracalla
yesterday')}) then this will be a phonemic distinction.  So long
as these forms remain in free variation, there is no phonemic
distinction.

>There was a whole lot more in Robert's last post, but it really
>just reiterates what has been said before.  What got my goat in
>his first post was the comment that (not quoting directly), "No
>one doubts that Modern English [th] and [dh] represent allophones
>of the same phoneme."

I'm sorry, John.  I didn't mean to get your goat, and now that I
have it I don't know what to do with it.  But since your
perception of what I said (as indicated by your paraphrase) and
what I actually said ("Most people would not insist on phonemic
status for both [th] and [dh] in English on the basis of this
minimal pair") have little relation to one another, I don't
consider that it is my fault.  I see a vast difference between
"No one doubts ..." and "Most people would not insist on ...".
So you can have your goat back with my compliments.  But if you
contintue to react to your perceptions of what is said rather
than what is actually said, then I predict that your goat will
continue to go walkabout with disturbing regularity.  I suggest
that you keep it on a shorter tether. :)

>I doubt it, and quite seriously.

I doubt it, too.  I just want to see some evidence that makes it
unequivocal.  Everything that I have found that *could* be taken
as evidence for [th] and [dh] being used as separate phonemes can
be convincingly explained otherwise.  I think it is something
that is in the process of happening and therefore hard to pin
down.  It is rather like the change in the meaning of 'fulsome'
that is currently taking place.  While both meanings are in use,
there is the danger of ambiguity.

>I also realize that there are multiple levels of "phonemic
>analysis" as represented by points of view ranging from pure SPE
>(where much emphasis is placed on the "native"-"nonnative"
>distinction between vocabulary) to the more structuralist
>approaches.  Perhaps when we respond to Pat and other
>non-professionals who occassionaly tug our chain, we can
>remember that professional linguists may all be walking in a
>westerly direction, but we're not necessarily arm-in-arm and
>keeping in step.  :)

Hear, hear.

Bob Whiting
whiting at cc.helsinki.fi



More information about the Indo-european mailing list