LINGUIST List 7.1231

Wed Sep 4 1996

Sum: /wo/ -> /jo/

Editor for this issue: Ann Dizdar <dizdartam2000.tamu.edu>

Directory

1. Bertinetto Pier Marco, Sum: /wo/ -> /jo/

Hallo everybody.

A couple of months ago I sent a message, concerning a sound change to
be observed in the Venetian dialect. Please note that "dialect" should
be intended here as "local vernacular", rather than "local variety of
the national language" (i.e. Italian). This is obvious for those who
are familiar with the sociolinguistic situation of Italy, but it might
be misunderstood by many.

Below I reproduce the original message. The colleague who addressed my
attention to this problem is Alfredo Stussi, with whom I discussed
this summary. We jointly thank all those who responded to our query.
Unfortunately, due to some clumsy manoever that I did in transfering
the messages from this address to my home address (and to Alfredo's
one), I lost part of the data concerning the contributors. Below is
what I could recover from my files:

Celso Alvarez-Caccamo, Universidade da Corunha (lxalvarzudc.es)
Paola Beninc`, Padova (benincaipdunidx.unipd.it)
Henno Brandsma, ??? (hsbrandcs.vu.nl)
Jakob Dempsey, Taiwan (???)
Richard Desrochers, Montreal <desrochrERE.UMontreal.CA)
John Dienhart, Odense (jmdlanguage.ou.dk)
Simon Donnelly, ??? (???)
Lance Eccles, Macquarie U., Australia (lance.ecclesmq.edu.au)
Thomas T. Field, University of Maryland (???)
James Gianola, ??? (jamessensoryc.com)
Ron Kuzar, Haifa U. (kuzarresearch.haifa.ac.il)
Mark Liberman, UPenn (mylunagi.cis.upenn.edu)
Michele Loporcaro, Z|rich (mloporcarom.unizh.ch)
Philippe Mennecier, LACITO-CNRS (Paris) (phmmnhn.fr)
Larry Trask, University of Sussex (larrytcogs.susx.ac.uk)
-s, ??? (Stavros Macrakis ???) (macrakisosf.org ???)

 * * *
 ORIGINAL MESSAGE 

On behalf of a colleague (but also out of personal curiosity), I would
propose the following phonological problem. A few words in the
Venetian dialect show the following change: /wo/ -> /jo/. Cf. (using
the Italian spelling): "siola" for "suola", "niora" for "nuora",
"frutariol" for "frutaruol" (It. fruttivendolo).

The questions are:
1) does anybody know of such examples in any other language?
2) What is the most likely explanation?

As to the latter point, the two proposals which have been put forth
are the following: a) dissimilation b) analogical attraction by the
far more frequent /je/ diphthong.
 * * *

To start with, we got the following query:

 * * * 

Are you sure there was a "change" from /wo/ -> /jo/? Perhaps /jo/ is
simply the Venetian reflex of the Vulgar Latin? -s

 * * *

This is not likely to be the case. The first, somewhat rare,
attestations of /wo/ -> /jo/ date from the XVIth cent., much later
than the normal /O/ -> /wo/ change.

 * * *

Several people noted that in the examples given in the original
message the diphthong was preceded by a coronal consonant. Thus, a
number of repondants (Desrochers, Dienhart, Donnelly, Eccles, Gianola,
Liberman, Loporcaro) asked whether this was fortuitous or not; in case
it is not, they suggest that the process is rather an assimilation
than a dissimilation. Beninc`, having a direct knowledge of the
linguistic data, noted that this is indeed the case, adding further
examples ("tjor" < "twor", It. "togliere"; "djol" < "dwol",
It. "duolo, dolore"), and remarking that this phenomenon (on the verge
of disappearing) is also to be observed, in isolated words, in some
Friulian dialects of Venetian influence. Among the most detailed
answers, M. Liberman proposed an "Ohalian" explanation:

 * * * 

First, given that the following vowel is rounded, a palatal on-glide
normally would be rounded to some extent as well, at least as a matter
of normal co-articulation. Thus the rounding feature of the original
/w/ has not been lost, at least phonetically speaking; the result is
the glide version of /y/, not of /i/. Second, the particular examples
that you cite all involve a preceding coronal consonant. Such
consonants commonly cause fronting of an adjacent /u/ at least to a
central and often to a front rounded vowel -- most dialects of
American English show this effect. If all of the examples of the
Venetian phenomenon are similar, then I would suggest that it starts
with a phonetic effect, in which a preceding coronal consonant fronts
a following /w/, resulting in a more-or-less front rounded
on-glide. Because an underlying /j/ would also normally be rounded
preceding /o/, listeners are free to misinterpret the resulting
pronunciation as originating from a lexical /j/ instead of a lexical
/w/. Given, as you suggest, that /j/ glide is more common than /w/ in
such contexts, then its greater frequency might well play a role in
generating the misinterpretation and in establishing it as the norm of
the Venetian speech community. This scenario is consistent with a
change that is lexically sporadic ("a few words...") rather than
phonologically regular. [...] I should also mention that the
proposed explanation, as you have probably noticed, is in the style of
John Ohala's suggestion for a general phonetic explanation of
dissimilation-as-falsely-assumed-assimilation. The idea is that
because of assimilation, a lexical sequence /...XZ.../ will tend
sometimes to sound like [...YZ...]. Listeners thus get in the habit of
interpreting tokens of the form [...YZ...] as representing the
phonetic reflex of /...XZ.../, even if the [Y] is actually the
phonetic reflex of a lexical /Y/. In the proposed explanation of the
Venetian case that you cited, the situation is a bit more
compicated---the idea is that in the environment between a coronal
consonant and /o/, both /w/ and /j/ can have a phonetic reflex that is
some kind of high frontish rounded glide. This makes it easy for
listeners to mistake the lexical 'spelling' of a particular word. In
keeping with Ohala's general approach to things, this attempts to
replace an entirely abstract cognitive event (dissimilation) with a
concrete physical one (co-articulation) plus a minimal psychological
construal of the physical facts (shift of phonemic category in some
words). The explanation seems to be quite reasonable in this
particular case.
 Mark Liberman
 * * *

Gianola suggested that underspecification and coronal as a default
place of articulation may have a role in this connection. This is of
course a plausibale, although theory-bounded view.

As to the first question (other languages?), we got the following
suggestions.

 * * *

Il existe un phe'nome`ne semblable dans certains dialectes inuit et en
particulier au Groenland oriental (EG). /ani-vu-q/ :
"sortir"-indicatif-3e personne, "il sort" Le mot est prononce', selon
les dialectes, (anivoq), (aniBoq), (aniwoq) et en EG (anijo)
 (entre () : les re'alisations phone'tiques ; B = v bilabial ; o
 pharyngalise')
De me^me, avec un radical en /u/ : /taku-vu-q/ "il (se) voit" -->
(takuvoq), (takuBoq) --> EG : (tagijo).
Le changement n'a pas lieu avec un radical en /a/ : /aala-vu-q/ "il
bouge" --> (aalavoq) --> EG : (a:laBo), (a:lao).
On peut supposer ceci : "suola" > (si-wola) > (sij-ola), avec
dissimilation de u > i.
 Philippe Mennecier

 * * *

The same phenomenon occurs in dialects of Occitan (notably those of
lower Languedoc). Latin o followed by a palatal consonant or a velar
stop: foliam > fuolha (pron. /jo/ in some dialects) 'leaf'; focum >
fuoc (/fjok/) 'fire'. By the way, other Occitan outcomes of this vowel
in these contexts = /wo/, /wE/, /j&/, sometimes with fronted /w/.

 Thomas T. Field

 * * * 

I know of one other example of this change, namely in eastern dialects
of (Westerlauwer) Frisian. (Spoken in the Dutch provence of
Friesland). There the standard Frisian wo is changed to jo, as is the
diphthong wa to ja. Examples include: /bjorn/ for /bworn/, spelling
buorren meaning "vilage center", /mjor/ for /mwor/, spelled muorre,
meaning "wall", and some more. This pronounciation also appears with
children in other parts of the provence, but there it is soon
corrected. The fact that in this eastern part this "childish"
pronounciation persists, is most often explained by the relative
difficulty of diphthongs starting in /w/, and the fact that many
non-Frisian settlers came to live in this area which had to learn the
language, and supposedly simplified the diphthongs. Also diphthongs
starting in /j/ are very common in Frisian, and have been in the
language for a longer time. Also, this seems to happen more after
labial consonants, so there might be a factor of dissimilation at work
as well, but that does not seem relevant to your Venetian examples.

 Henno Brandsma
 * * *

The last contribution is intriguing, for it shows that this particular
change needs not be a case of assimilation, since the preceding
context is not necessarily coronal. Indeed, dissimilation seems to be
at stake in other cases as well:

 * * *

(in case it helps) the /ow/ ~ /oj/ alternation is frequent in
Galizan-Portuguese: cousa ~ coisa, ouro ~ oiro, pousar ~ poisar, etc.
/oj/ is a later development, as etymologically the diphthong is /ow/:
CAUSA > cousa. [What is the most likely explanation?] Probably
dissimilation. Also Old Spanish (and Astur-Leonese) /wo/ -> /we/:
PORTA > poOrta > pworta > pwerta In this case there was probably
analogical contagion by /je/ : PETRA > pEdra > pjedra
 Celso Alvarez-Caccamo

 * * *

The parallel is not perfect, but eastern varieties of Basque exhibit
something similar to the change you're interested in. The sequence in
question here is /ua/, since /uo/ hardly ever arises in Basque. The
Basque definite article is a suffix /-a/, as in <gizon> `man',
<gizona> `the man'. Since Basque is not fond of vowels in hiatus, the
addition of the article to a vowel-final noun often results in some
kind of phonological process which has the effect of making the hiatus
less severe; the details vary considerably in the several dialects of
Basque. In Labourdin and Low Navarrese, a glide [j] is inserted.
Thus, with <esku> `hand', whose definite form is written <eskua>, we
find [eSkuja] in L and LN (here [S] represents the Basque apical
sibilant, notated <s>, which contrasts with the laminal sibilant [s],
notated <z>), while <burua> `the head' comes out as [buruja] (with a
tapped [r]). Some Pyrenean varieties reportedly have [burja] for this
last, and similarly for other words, but syncope is typical of these
varieties. Farther east we find something different. In the Souletin
dialect, historical /u/ is fronted to /u"/ (phonetically [Y], as in
German <fu"nf> `five') in most circumstances (the fronting is blocked
by certain neighbouring sounds). Thus <esku> `hand' is [eSku"] here.
But the definite form <eskua> appears as [eSkia] -- that is, the [u"]
is unrounded to [i] before /a/. Note, though, that this last form
remains trisyllabic: there is no glide-formation. In the recently
extinct Roncalese, just south of Souletin, /u/ has not undergone
fronting in general, and hence we have [eSku] for `hand', but `the
hand' is none the less [eskia], and similarly for other such words.
That is, /u/ is fronted to /i/ before /a/, though again it apparently
remains syllabic. So, we have two different developments: (1) [ua] >
[uja] ( > [ja]) (2) [ua] > [u"a] > [ia]

 Larry Trask

 * * *

Kuzar pointed out a different, but somehow similar phenomenon:

 * * *

In Hebrew there is often a change from uo or oo to io, or from uu or
ou to iu, except that it happens with a consonant in the middle:
uCo>iCo. -on or -oni is a suffix that adjectivizes nouns: Examples:
roS 'head'. roS+on=riSon 'first, principal' xuc 'outside'
(noun). xuc+on/oni=xicon/xiconi 'outer' xol (stem: xull-) 'profane'
(noun: 'the profane') xull+oni=xilloni
 'secular' 'profane' (adjective)
Finally: the adjectives of the structure XaXuX become abstract nouns
by the addition of -ut='-ness', but the form is XXiXut (the zeroing of
the a is an unrelated process). Example: naxuc 'necessary' nxicut
'necessity'. The traditional explanation is dissimilation.
 Ron Kuzar
 * * *

Finally, and happily, Dempsey pointed out that the reverse change is
also documented:

 * * *

Actually, I cant think of any examples offhand, but I know of a
Chinese dialect (Fuzhou) where the OPPOSITE happens: -jo > -wo !
 Jakob Dempsey
 * * *

We hope this can be of some utility. Thanks again to those who
provided informations.

Pier Marco Bertinetto & Alfredo Stussi
Scuola Normale Superiore
Pisa




 ||||
 Pier Marco Bertinetto ____
 ///////
 tel. ++39/(0)50/509111 -------
 fax: ++39/(0)50/563513 ///////
 -------
 ///////
 ------- Scuola Normale Superiore
 /////// p.za dei Cavalieri 7
 ------- I-56126 PISA
 ///////
 =====================================================