LINGUIST List 9.991

Wed Jul 1 1998

Sum: [ae]>[ya] summary

Editor for this issue: Brett Churchill <>


  1. WEN-CHAO LI, [ae]>[ya] summary

Message 1: [ae]>[ya] summary

Date: Tue, 30 Jun 1998 19:01:29
From: WEN-CHAO LI <>
Subject: [ae]>[ya] summary

[ae]>[ya] SUMMARY By Wen-Chao Li

A while ago I posted a querie on the list asking for information on
strategies used by languages that do not have the low front vowel [ae]
when they adapt loanwords with the vowel [ae] from foreign languages.
I received a number of very helpful replies: thanks to Robin Thelwell,
John E Koontz, Jakob Demptsey, Colin Whiteley, Yehuda N. Falk, Daniel
E. Collins, Jeroen Wiedenhof, Bob Allen,Markus Hiller, Ivan A
Derzhanski, Laurie Bauer, and anyone else that I might have left out.

(1) THE [ae]>[e] POSSIBILITY

In my original querie, I mentioned that I knew of two common strategies for
adapting the vowel [ae]: 

(1) [ae]>[a]
(2) [ae]>[ya]

Many respondents were quick to point out that there is a third
strategy that is even more common, namely, [ae]>[e]. Thus Daniel E
Collins <> writes that

"In the Slavic languages best known to me, English [ae] is often
borrowed as [e] (e.g., Russian "kemping"/camping, or the pronunciation
of my name as "den"/Dan)".

Laurie Bauer <>: "English loan words in German
replace the [ae] with open E (CV3)".

Jeroen Wiedenhof <>: "In Dutch, which does
not have a phoneme /ae/, the English phonemes /ae/ and /e/ in loans
are *both* reflected as the Dutch phoneme /e/ (=IPA epsilon sign).
Compare the following Englishloans in Dutch:

"camping" ['kempIG]
"petticoat" ['peti.kot]
"patchwork" ['petSV$Wk]

['=primary stress; .=secondary stress; G=ng; S=voiceless postalveolar
 fricative; V=voiced labiodental fricative; $=rounded open-mid vowel
="oe" ligature; W=voiced velar approximant]"

Yehuda N. Falk <> gives examples from Hebrew:
"I did want to point out that there is a third possible treatment of
[ae]: it can become a mid vowel [e] (or epsilon). This is what
happens in Hebrew.

kitbag kitbeg
tramp tremp (means "a ride; a lift")
cat a brand of catfood called "ketli"
handout hendawt (heard at linguistics conferences)
captain kepten

(2) [ae]>[ya] AFTER VELARS 
Of the three available options in the adaptation of loanword [ae], I
was most interested in the [ae]>[ya] phenomenon. In my original
querie, I gave Japanese as an example of [ae]>[ya] re-analysis, and
cited the following examples:

"camping" kyanpingu
"gallery" gyarari
"gang" gyangu
"cabin" kyabin

Jamaican creole behaves in a similar manner:

"can" kyan
"can't" kyaan
"carry" kyai

What bothered me about these examples, however, is that they seemed to
occur exclusively in velar environments. If that is the case, then
velar palatalization would seem like a more likely explanation than
re-analysis of the [ae] vowel. One of the purposes of my querie was
to see if there are languages in which the [ae]>[ya] process occurs in
non-velar environments.

Not surprisingly, many respondents gave examples of [ae]>[ya] following

Robin Thelwall <> writes that "palatalization of velars
before [ae] is also found in Ulster Scots dialects of English". 

Bob Allen <>: "The same phenomenon appears in the Welsh spoken
around Bangor, North Wales, and possibly in other Welsh dialects. As in your
Japanese examples, it is restricted to velars ... The written source for this
information is Fines-Clinton (sorry I don't have the first name): WELSH
VOCABULARY IN THE BANGOR DISTRICT, which was published back in the 1920s or

John E Koontz <> writes of a Hank Williams Sr song in
which "care" is pronounced as [kyar]. "I'm not sure where Hank Williams was
raised. Somewhere in the southern tier -- Arkansas or Oklahoma?"


But in the end, someone did come up with examples of [ae]>[ya] in non-velar
environments. This occurs mostly in the Slavic languages. 

Colin Whiteley <> writes that "Russian is a
good candidate, since it has a whole series of contrasts
palatalised/not palatalised, and so it must choose which to take when
the original language has an intermediate value. So "beach" =
"plyash" from French "plage". However, the French /a/ is a lower
vowel than you describe and the point in this case is
that the French "l" sounds closer to the Russian palatalised "l"
than to the "dark" one (in Russian they're described as "soft" and

And, saving the best for last, Ivan A Derzhanski's <>
huge list of examples:

"The standard rendering of Finnish and Estonian names in Russian and
Bulgarian prescribes the substitution of "ya" (ie /a/ preceded by
either /j/ or palatalisation) for the vowel [ae]. In a few cases I've
also seen the English /ae/ rendered as "ya" in Bulgarian -- in names
of football teams, if family serves.

Some examples of Finnish in Russian: 

Juvjaskjulja Jyvskyl (a city)
Pjajjanne Pijnne (a lake)
Vjajnjamjnen Vinminen (one of the main characters in the 

I use "ju" and "ja" for the last two letters of the Cyrillic alphabet,
which stand for palatalisation or /j/ before /u/ and /a/, respectively; ""
the Russian letter "e" - diresis (palatalisation or /j/ followed by /o/); "j"
for the Cyrillic "i" - breve (/j/); and "'" for the soft sign

Note that "" and "y" are rendered as "" and "ju" (so in this the treatment
consistent: any /{+frt,+vwl}/ which does not exist in Russian is divided into
/{+frt}/ (rendered as /j/ or palatalisation) and the rest of its features
(rendered as a back vowel). Note also that "j" before "", "" or "y" is lost.

Estonian works in the same way, e.g., 

PjarnuPrnu (a river and a port at the Riga Bay)
mjagi mgi ("mountain" -- in many place names)

Bulgarian works in a very similar way, except that there is no letter "",

Vjajnjamjojnen Vinminen (a purely graphical difference) 

and the sequence "jja" is considered less acceptable, hence 


which looks as if the "i" were syllabic, although in fact it is not;
apparently they were loath to let the second "ja" do the job of three
(!) separate facts of the Finnish -- (1) the coda of the preceding
syllable, (2) the onset of this one, and (3) the frontness of the

Found a bunch of interesting info in Andrej Danchev (1978), "Bulgarian
Transcription of English Names: Theory and Practice" (the book itself
is in Bulgarian; it just has a bilingual title).

The story is as follows. The standard rendering of English // in
Bulgarian used to be "e"; now "a" is taking over, reflecting a
defronting of the corresponding vowel in English itself.

However, in some older sources "ja" is also found, and things like
"Bljak"/Black, "Njashnl siti bjank"/National City Bank ("" meaning the
schwa vowel, the Russian hard sign) have appeared in print.

I myself have seen "Sautxjamptn"/Southhampton (a football club).

AD notes that this transcription is looked down upon, although it is
no different from the substitution of /ju/ for /y/, which is standard
in the transcription of German, French, Turkish etc. names, and which
led to the current pronunciation of words like "future" in English;
nor is it alien to the history of English, where (in Early Modern
English) spellings such as _cyan_ for _can_ have been attested (an
attempt to represent the fronting). The only case in which "ja" for
English // has become standard in Bulgarian (as well as Russian) is
"Aljaska"/Alaska (which afaik no one has attempted to represent

AD also notes that "ja" is (or was) more readily used after /l/, /k/,
/g/ and /h/ (the last rendered as "x" /x/), that is, the consonants
which are affected to the higher degree by a following front vowel; it
is never seen in word-initial position.

Finally, "ja" tends to be seen in transcriptions made by speakers of
eastern Bulgarian dialects, in which the Old Slavic // yields "ja";
those from the west, where OSl // became "e", are more inclined to use
"e" for the English //."


Finally, some additional comments:

Markus Hiller <> adds that "Japanese
adapts German front rounded vowels in a similar way, e.g.

<Muenchen> ['mync,n] > [mjunhen] (place name). 

As you can see from the example, this is *not* limited to velars, but
the mid front rounded vowel is also often adapted as [e]. Many slavic
languages do virtually consistently adapt German front rounded vowels
as palatalization
back vowel, i.e. 

"C+[y]" as "palatalized-C+[u]" 
 ("y" here is rounded high front vowel)

<Muenchen> [mjunxen] (see above) 
<Koeln> [k0ln] [kjoln](place name) 

If you are interested in decomposition, you might also be interested
in looking at these data: many colloquial varieties of northern German
adapt French [,n] (palatal nasal) as [Nn] (velar nasal plus alveolar
nasal), whereas standard german has [gn] in that place. Thus you get
[maN'ne:t]~[mag'ne:t] "magnet", whereas [mag'na:t] "tycoon" comes from
a different source and never is *[maN'na:t]. (Strangely, <cognak> is
['kOnjak], not *['kONnak]). Hall (1992) analyzes /gn/ to be underlying
both in [maN'ne:t] and in [mag'na:t], which does not look like a good
solution to me."

Jakob Dempsey <>: "The change [e]>[ya] is rather common
in diachronic development, but not [ae]>[ya], unless the [ae] were interpreted
as a low [e]".

Compiled by 

Wen-Chao Li, Lecturer
Institute of Linguistics & Asian & Slavic Languages & Literatures
University of Minnesota

Get free e-mail and a permanent address at
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue