LINGUIST List 12.57

Thu Jan 11 2001

Sum: Learning Trills

Editor for this issue: Lydia Grebenyova <>


  1. Julian Bradfield, Learning Trills

Message 1: Learning Trills

Date: Fri, 5 Jan 2001 12:18:08 +0000 (GMT)
From: Julian Bradfield <>
Subject: Learning Trills

More than a year ago, I posted (10.1160) a request for any hints on
learning trills, and as a subsidiary question, whether there is in
fact any widespread physical inability to produce any common sound
(something asserted about [r] in an earlier discussion). Over the
following month or two, I received several suggestions and comments.

I apologize to all those who were waiting for the summary; I hope that
it may be still useful, and at least it will be in the archives for
future searches.
(This message is Bcc'ed to all those who sent me comments.)

First, the question of practical advice on learning to pronounce [r].
I got many suggestions on this, many of which were along the lines
"repeat taps quickly and relaxedly". I can't actually say that any of
the suggestions was the magic bullet for me; over the year since I
posted, my [r] production has improved, to the point where in normal
use it's probably OK, but I still can't sustain it for more than a few
beats. Perhaps the most useful comment for me was "remember that [r]
is a very forceful sound" (i.e. don't expect to practise without
disturbing people!). Here are the suggestions:

Kathleen Tacelosky:
A trick I use with my students (mostly native English speaking =
undergrads.) for teaching them how to feel the position of the tongue is =
to say "I edited it" really fast. I think it was a professor of mine who =
taught me this. 

John E. Koontz:
Try it voiceless in final position. Try switching between uvular and apical. 
For what it's worth, trilled r is very forceful. My vision actually jumps
when doing it. 

Mandy Schiffrin:
[ a phonetics teacher said that ]
you have to train the muscles in the mouth which would be 
developed as a matter of course in speakers of languages where the [r] 
occurs "naturally". You do this by repeating the phonemes [t] and [d] 
(with some kind of neutral schwa sound in between) as fast as you can, 
for say five minutes a day. Eventually the trill becomes easier as you 
train yourself. It worked for a friend of mine...

Rob Pensalfini:
I don't have much advice on producing trilled /r/, though I have succeeded
on two occasions in teaching people to. Many people seem to find it easier
to learn to procuce the voiceless variant, and for some reason find adding
the voicing in hard, but once they master the voiceless trill, can add
voicing with time.
[ and in a subsequent conversation: ]
One of my non-linguist hats is actually as a voice teacher for actors, and
I am trained in the Linklater tradition. There, the received wisdom is that
phenomena such as short tongue, narrow tongue, crinkly tongue et cetera are
not actually genetic (or at least not entirely) but result from habits in
the use of the speech organs, usually relating to storing tension or
'holding' with the muscles of the tongue, soft palate, jaw et cetera. The
Linklater technique involves a number of exercises for loosening these
various organs, including a number of tongue stretches and shakes, which I
have found incredibly useful to the production of clear sound and
articulation alike.

Jorge Guitart:
Don't know if this works but try to put your tongue where you would for
English /l/ and think /tr/, then get rid of t later on.
Let me know.
If you were trying to do a spanish trill I would advise you to subsitute
the strident r used in the Andean region of Ecuador and in Chile.

C. Whiteley:
I am exactly the same sort of sufferer that you are and have had
problems with [r] all my life. It took me years to come to grips (more
or less) with the various allophones of r in English, and I still have
problems with a sustained trill in Spanish, which as I'm sure you know
is phonemically distinct from the single tap. Some native Spanish
speakers have a similar difficulty, so my Spanish born wife and
children tell me my r sounds oddly Spanish rather than merely foreign.
My approximation to a trill is in fact a series of only 3 or 4 taps,
which to a Spanish ear sounds as though it is preceded by a /d/ (a
labio-dental fricative/approximant. I know that my tongue tip should
not touch my teeth at all but should vibrate loosely in the air stream
without moving either forwards or backwards. My attempts at achieving
this are pathetic!
I wonder whether difficulties with the trilled r are related to the
better known difficulty many English speakers have with the standard
pre-vocalic and intervocalic approximant, often replaced for example
with a labio-dental approximant.

The second question was whether there exists any widespread physical
disability to produce any common sound, and [r] in particular.
Responses varied from "yes, but probably not very significant" to
"stuff and nonsense". Here are the responses (ignoring comments about
obvious deformities such as cleft palates):

Norvin Richards: 
Don't know how significant the numbers are, but there apparently are such 
people, yes; I remember hearing that there was an Italian linguist who was 
such a person (Gennaro Chierchia, maybe?), which suggests that it isn't just 
a matter of lack of exposure to the sound during the critical period.

Rob Pensalfini:
I am really writing about your other question, the issue of 'sub-standard
mouths'. I do suspect perhaps that some of the teachers are 'excusing'
themselves for their failure at teaching the trill, however, my
undergraduate syntax teacher claimed that he could not trill /r/ because he
had no tongue tip. He claimed to have been born without one. I don't know
if it's true, I guess it's certainly possible, but I don't know how common.
[ and see above also ]

Peter T. Daniels:
This was a commoplace assertion by Japanese linguists of the
fascist/racial superiority era and is still sometimes heard from
unsophisticated Japanese persons.

Mandy Schiffrin mentioned being "tongue-tied" (ankyloglossia
inferior), which is a not uncommon problem, and can affect a number of

Anthea Gupta:
There is no widespread physical ability to produce any sound used in 
any human language. There are vocal tract problems that prevent 
particular sounds being made (e.g. if you have no vocal chords you 
can't produce voiced sounds; if you have a cavity in the roof of your 
mouth you can't produce non-nasal sounds; if you don't have any front 
upper teeth you can't produce dental fricatives). Most people who 
have a physical problem that prevents them from making language 
sounds know about it.
One example I always use with my students is that of tongue grooving. 
 This is a genetically determined ability which is present in about 
half of most populations (making your tongue into a U with the sides 
up). With the result that no human community requires it in the
phonology of a language. If you can groove your tongue, you can push 
air through it and get a sound very much like a dental fricative (I 
always have to get the students to demonstrate this, as I can't do it 
myself). So a tongue-groover with no front teeth could produce a 
better quality of speech than I could if my crowns had to go.....

Johnny Thomsen:
As to your sub-question, I agree with you that it is quite improbable that 
there should be any widespread physical inability to produce any common 
On the other hand, in languages with a 'trilling' *r* or something close to 
it, the inability to pronounce it seems to be found in varying degrees.
In my native Faroese, which has a weak trill in its *r*, I have never met 
this phenomenon myself, but I have heard and read about its existence. It 
seems it was locally restricted, even to some individuals. There is even a 
special Faroese verb for this, "skurra" or, more rarely, "skarra". The verb 
is, though, unfamiliar to most Faroese, so the phenomenon cannot have been 
In Russian, which has a real trill in its *r*, the verb for this is 
'kartavit', and the verb is well known among Russians. I have heard this 
manner of pronouncing Russian myself, and I know it is, if not widespread, 
at any rate sufficiently frequent. The most famous example of this defect 
was Lenin himself, which is rather unfortunate when you are the leader of 
Rossiyskaya sotsial-demokraticheskaya rabochaya partiya (the Russian Social 
Democratic Workers' Party).
If I am not mistaken, the trilling *r* is one of the last sounds children 
learn in native language acquisition. I know for sure that Faroese children 
sometimes pronounce a /j/ sound instead of /r/ long after they master all 
the other sounds of the language.

John Koontz:
I wonder if there aren't some people who just have extreme difficulty with
the sound, though? A number of Americans can't manage the retroflex r and
substitute various things for it, e.g., w, as lampooned in the cartoon
character Elmer Fudd. I wonder if there aren't similar problems in
languages that have apical trills. Something like that might explain the
switch to the easier (I think) uvular trill and fricative common in much
of Europe. 
There are also two competing s's in American English. Some people can
manage the retracted version better than the apical one. Other people
consider the retracted version to be a speech defect.

Other similar comments were received from
Deborah Milam Berkley
C. Whiteley
Philippe Mennecier
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue