Re: Languages and Timbre

Subject: Re: Languages and Timbre
From: Kevin Austin (
Date: Sat Nov 06 2004 - 22:56:39 EST

 From my understanding, the two biggest impediments to speaking
without an accent are "hearing" and conditioning.

The vocal tract system adapts itself to 'favor' the production of a
limited number of the sounds possible (and transitions between these

I will return to the 'hearing' a little further on.

For 'favored sounds' (which includes elements of segmentation),
consider the Polish letter group "szcz". For english speakers this at
first is not easy to grasp as it is not a 'segmented group' in
english. It does occur as a sound. Say the phrase "Wish children" ...
the 'szcz' sound is that of "sh-ch" in 'wish children'. Or try 'Krz',
which is "ck-sh" from the phrase "kick shoes. The individual sounds
exist, but not as they are used in the other language.

That's a simple case. A more complex one is that of sounds which are
not found in the second language. "This thing" will often be
pronounced "Dis ting" by non-english speakers as they have not
learned the "th" sounds (voiced and unvoiced). Instead, the speaker
substitutes the (only) sound which their physiology has comfort with,
in this case, the labial-dental consonant(s) /d/ and /t/.

Once the speaker has learned how to allow air past the sides of the
tongue, the /d/ and /t/ become the two forms of /th/.

And there are more complex examples, such as the 'tongue click' found
in zulu and other south african languages.

None of this can be achieved until the ear "hears" the difference.
English has (basically) one /l/ sound ... the tongue goes to the roof
of the mouth, but some southern Indian languages have as many as
three or four different /l/ (like) sounds. The english (or french)
listener has to be trained to differentiate the differences, and then
to produce them. Not impossible, but potentially challenging.

Another interesting case is that of japanese which has a liquid
consonant (about) halfway between /l/ and /r/. While english has two
(different) sounds, the japanese one is about 'halfway' in between.

When an untrained japanese person says "flight", many english
speakers (think they) hear "fright". This is because the /l/ sound
appears more /r/ than expected.

Parallel, when the japanese native says "free", many english speakers
hear "flee:, as the /r/ sound appears more /l/ than expected.

In many cases both of these sounds (the japanese person saying /l/
and /r/) are the same. The japanese speaker does not (yet)
differentiate between the /r/ and the /l/.

Similarly, the english speaker hears that the (expected) /l/ sound is
too /r/ -like, and the (expected) /r/ is too /l/-like, and thus
'hears' the 'other' sound (or rather interprets it as such).

The best non-native speakers of Polish are the Italians. It just
happens that the 'value' (formant structure) of almost all of the
vowels in the two languages are 'identical'. Polish instructors have
told me that most (native) italians who learn Polish without an
accent as far as the voels are concerned.

The situation is added to with two other aspects, the stress / length
of syllables between languages -- english is based upon 'stress'
created by making some vowels stronger (louder), while french stress
is based upon the length of the vowel. An english speaker tends to
"accent" french words rather than changing the speed of the vowels; a
french speaker tends to 'flatten' the stress patterns of english
words and transfers the patterns which they know, onto the new

The other aspect is that of phrase / sentence long 'intonation
curves', a characteristic not found in all languages.

Ask this question: "Which way did he go?" There is a strong tendency
for the pitch to rise at the end, indicating a question. In chinese a
question is not indicated by a rising intonation curve ... that could
change the words which are in the sentence ... but rather by the
addition of the word "ma".

Are there innate physiological differences? A quick look at newscasts
from around the world may indicate that (in general), not. A few
years ago, I listened to a Scottish newscaster speaking in a very
thick Scottish brogue (accent). When I looked, I was (very) surprised
to see a black woman whose parents had probably emigrated to the UK
40 years earlier. The young black lady had not a hint of "islands"
accent, and had completely adopted a Scottish accent and mode of

(There also a number of other psycholinguistic and psycho-social
considerations that appear to be determined more by environment than
by genetics.)

The vocal cords (not chords) do not 'produce' formants. Formant
structures are produced by the vocal tract.

I appear to have an english accent in Canada, in the States I have a
Canadian accent, and in the UK, I tend to sound like a mixture of
Canadian and American.

Go figger eh!



At 20:01 -0500 2004/11/06, Jason Smalridge wrote:
>I would like some help settling a discussion about timbre in a voice
>made from different languages. I believe that the voice is created
>by lungs, vocal cords, tongue, teeth, lips... and that each language
>has a distinct way of using these parts to create the timbre of
>their language. I thought that this was one reason why if a person
>who's mother tongue is English began to learn Japanese, his accent
>would be noticeable until he learned how to create the specific
>timbre to speak the language.
>I know that timbre is sort of hard to define, but I guess what I am
>asking is: Can a person who speaks two languages perfectly(without
>a trace of accent) actually be defined as a person who speaks with
>two different timbres? Is there more at play then just the way the
>lips and tongue form the different words from different languages?
>Do the vocal chords actually change the (I think their called)
>formants to achieve different languages?
>Blah blah blah....

This archive was generated by hypermail 2b27 : Sat Dec 22 2007 - 01:46:04 EST