Re: Subtitutes for English /T/ and /D/

In article <sfg4a3t5a8qjjqs6idrgccqtk9m1n70lab@xxxxxxx>,
Ruud Harmsen <realemailonsite@xxxxxxxxxxxxxxxxxx> wrote:

Sat, 21 Jul 2007 11:24:06 -0400: Nathan Sanders
<nsanders@xxxxxxxxxxxx>: in sci.lang:

Phonemes either are or are not.

I objected to that earlier. If that were so, how can [f] and [v] have
been allophones in Old-English and have developed into separate
phonemes in Modern English?

When the trigger conditions for allophony were eradicated.

If your statement were true, there must have been a fixed date at
which such changes took place.

The date when a child acquires a language that needs them to be
phonemes, if any such date can be determined.

Perhaps not the same date for all
speakers (and each potential phoneme pair), but one determinable date
for any given speaker then alive. Or was there a period in which, for
one speaker under consideration, in 70% of his utterances the sounds
were separate phonemes, and 30% allophones? More or less by random
choice? And a year later this had become 72 and 28%?

As long as the allophones are derived synchronically, then allophones
they remain, no matter how (in)frequent the rule is.

Suppose we have a language with an English-like history, spoken by a
man we shall call Bob. In Bob's idiolect, nouns form their plurals by
regular suffixation of unstressed /-as/, and any stem final fricatives
are voiced when they occur between vowels. Thus, the noun /tof/ would
be pronounced [tof] in the singular and ['tovas] in the plural.

Bob feels the urge to syncopate unstressed vowels, and slowly begins
to drop the unstressed [a] from ['tovas]. It begins small, only once
in a hundred times does he say [tovz] (with voicing assimilation), but
as his life goes on, Bob says [tovz] more and more frequently, until
by the end of his life, he never uses the older ['tovas] anymore,
because his syncope rule has reached 100% application.

Throughout all of this time, his phoneme inventory hasn't changed. He
just added a new rule to his grammar, whose chance of applying
increased over time. But the phonemic representation of [tovz] always
remained /tof+as/.

Bob's daughter Carol will acquire her language (partly) from Bob's
speech. Even if every instance of the language she hears from every
speaker only uses [tovz], we could still describe her grammar without
needing phonemic /v/ in this word, because [tovz] can be derived from
/f/ with application of two ordered rules: intervocalic voicing and
syncope. If every instance of [v] in Carol's speech fits into this
category (predictably derived by phonological rules from /f/), then we
would have no need to posit /v/ for her idiolect at all.

Since phonemes are only theoretical abstractions, this is all that
matters. Until we can directly examine speaker's brains, we can't
know for sure whether phonemes even exist as such. It's very likely
that all we'll ultimately find is something more like a neural
network, with nothing to point to that we could reasonably call "/f/"
or "/v/".

Same question for separate phonemes merging into one, in the history
of other languages.

Same answer: the rule that causes the merger varies in frequency over
time, until it reaches 100%. During that time, a description of the
language would note that a sound change is in progress, and thus, that
there are two separate phonemes. If the description happened to be
made after the rule reached 100% frequency for every speaker, then the
description would posit only a single phoneme (assuming of course that
this was a full, unconditioned merger, leaving no telltale remnants of
the former pre-merge contrast).

Nathan

--
Nathan Sanders
Linguistics Program
Williams College
http://wso.williams.edu/~nsanders/
.

Relevant Pages

  • Re: Your first "linguistic" memory
    ... >> They are different all right, but is FAL really a Dutch word? ... >> using that gun ever use it. ... They're not allophones either, because the difference can ... Phonemes are what make words sound different, ...
    (sci.lang)
  • Re: The "u" and "v" in older written English is confvsing
    ... phonemes have been merging into one phoneme with allophones in ... the notion that scripts don't determine pronunciation. ...
    (sci.lang)
  • Re: The "u" and "v" in older written English is confvsing
    ... phonemes have been merging into one phoneme with allophones in ... the notion that scripts don't determine pronunciation. ... brought about by the influence of script would be interesting. ...
    (sci.lang)
  • Re: Subtitutes for English /T/ and /D/
    ... been allophones in Old-English and have developed into separate ... Same question for separate phonemes merging into one, ... of other languages. ...
    (sci.lang)
  • Re: Latin pronunciation puzzle
    ... Faze with a rather soft f, as the accentuation is on z. ... Flocks with f, ... Here I don't rely on language I hear, ... a sonnet by Shakespeaere using the regular phonemes, ...
    (sci.lang)