Byrd - An Examination of PIE Ejectives to Voiced Stops in Italic Using Features (2013).pdf

(104 KB) Pobierz
Arrick 1
Darin Arrick
Dr. Andrew Byrd
LIN 515 – Phonology
1 May 2013
An Examination of PIE Ejectives to Voiced Stops in Italic Using Features
I
NTRODUCTION
Because Proto-Indo-European (PIE) is an unattested language, we have no direct evidence of its
phonological structure. Therefore, we use attested languages to reconstruct the possible phonology of
that parent language using the comparative method. But that method isn't perfect, and multiple
competing theories of the phonological system of PIE exist. The older, more traditional, theories posit
three or four series of stop consonants in PIE: unvoiced unaspirated, voiced unaspirated, voiced
aspirated, and sometimes unvoiced aspirated. These are spread across five places of articulation: labial,
dental, palatovelar (or velar), velar (or uvular), and labiovelar. While there are symmetries to be found
in these systems, some theories seem to lack a sort of typological symmetry normally seen in language
(e.g. lacking unvoiced unaspirated stops when voiced aspirated stops exist in the inventory is
extremely rare). Therefore, other theories for the reconstruction of the stop system of PIE have been
proposed. One of those theories is the "Glottalic Theory", which proposes reconstructing the voiced
unaspirated series (often called Series I) with unvoiced ejective stops instead. This fixes some of the
symmetry issues, but raises others. This paper will model the possible paths of change from these
reconstructions of PIE to the Italic languages (specifically Latin) using features to attempt to determine
the most-possible path and, hopefully, shed some light on the Glottalic Theory in general.
D
ATA
As mentioned above, the traditional reconstructions of PIE include three or four series of stops
over four or five places of articulation (the palatals/palatoalveolars won't be discussed here as they are
Arrick 2
usually the questionable place). The most-commonly accepted theory posits three series of stops. The
unvoiced unaspirated stops include /p/, /t/, /k/, and /kʷ/. The voiced unaspirated stops include /b/, /d/,
/g/, and /gʷ/. And the voiced aspirated stops include /bʰ/, /dʰ/, /gʰ/, and /gʷʰ/.
This three-series reconstruction is disliked for a few reasons. Mainly, it is typologically
unusual. When a language has voiced aspirated stops, it seems to always have unvoiced aspirated stops
as well (which, themselves, are more common than the voiced aspirates across languages). Also, the
bilabial voiced unaspirated stop, /b/, is rare in reconstructed PIE words, causing some confusion due to
its relative absence. It is because of these issues (amongst others) that the Glottalic Theory was
originally proposed.
The Glottalic Theory was developed independently by Thomas Gramkrelidze (Georgia) and
Paul Hopper (US). It posits that the stop series which, in traditional reconstructions, contains voiced
unaspirated stops (i.e. /b/, /d/, /g/, and /gʷ/) actually contained ejective stops (i.e. /p’/, /t’/, /k’/, and
/kʷ’/), giving a three-series stop inventory of unvoiced stops, ejective stops (which are, by nature,
unvoiced), and voiced stops (with possible phonetic variation in aspiration). This brought a sort of
symmetry to the system, giving a three-way manner distinction that is typologically more common.
The traditional reconstruction is held by most Indo-Europeanists, but there are linguists who hold to
the Glottalic Theory. As ejectives are often not studied by Indo-European linguists, a bit of background
would be helpful.
Ejectives are a class of sounds which are created by increasing pressure in the oral cavity
during a stop's closure by raising the larynx, causing a strong burst of air when the closure is released.
Ejectives are unusual in that they are the only class of sounds which are articulated by specific
movements of the larynx up or down, and they are always doubly-articulated; that is, they have a stop
closure elsewhere in the oral tract as well as at the glottis. These sounds are found in about 15% of the
languages of the world (Maddieson, sect. 2). According to WALS (the World Atlas of Language
Arrick 3
Structures), the only languages in Europe or Asia which contain ejectives are in the Caucasus.
Languages from this region have existed there for thousands of years; they are indigenous to the area,
so far as we know. It isn't unreasonable to think that they came into contact with PIE at some point.
One might even suggest that, based on the various theories concerning the PIE homeland, ejectives
were an areal feature of eastern Europe and southwestern Asia.
For the rest of the paper, the terms "traditional" or "glottalic" will be used to refer to the two
sets of sounds reconstructed for Series I of PIE. The feature matrices for the phonemes in the
traditional reconstruction are as follows:
/b/
+consonantal
-sonorant
-continuant
-delayed release
-approximant
+voice
-spread glottis
-constricted glottis
+labial
-round
-coronal
-dorsal
+consonantal
-sonorant
-continuant
-delayed release
-approximant
+voice
-spread glottis
-constricted glottis
-labial
-round
+coronal
+anterior
-distributed
-strident
-dorsal
/d/
Arrick 4
/g/
+consonantal
-sonorant
-continuant
-delayed release
-approximant
+voice
-spread glottis
-constricted glottis
-labial
-round
+dorsal
+high
-low
/gʷ/
+consonantal
-sonorant
-continuant
-delayed release
-approximant
+voice
-spread glottis
-constricted glottis
+labial
+round
+dorsal
+high
-low
The first five features in each list simply define them as stop consonants. They are all voiced. The
glottis is neither open nor closed, allowing vibration for voicing (these feature values are somewhat
redundant, but they will be used later for the glottalics so they are included here). The remaining
features are place features and lip roundness. These features are more than enough to describe these
sounds and allow us to discuss any possible changes.
According to the feature model used in Hayes (99), the glottalics vary from these four sounds
by only two features: {+constricted glottis} and {-voice}. But changing only these two features alone
is not enough to describe the glottalics, as the movement of the larynx is not considered. Avery and
Arrick 5
Isardi suggest a hierarchical organization of the laryngeal node in feature geometry to account for three
different aspects of laryngeal action. Hayes only takes into account what Avery and Isardi call Glottal
Width. This encompasses two features which describe the width of the glottal opening: {±spread} and
{±constricted}; if both are negative, the glottis is in a relaxed state. Glottal Tension describes the
tenseness of the glottal folds themselves: {±stiff} and {±slack}. These can be used to define breathy
and creaky sounds. The third variable is Glottal Height, the height of the larynx in relation to its neutral
rest position, and is described with two features: {±raised} and {±lowered}. These can be combined,
both with negative values, to refer to the rest position (similar to how {-high, -low} vowels are mid
vowels). In this model, an ejective would not only be {+constricted glottis}, but would also be
{+raised}. This describes the act of closing the glottis to prevent airflow back into the lungs and raising
the larynx to increase the intraoral pressure, the defining trait of an ejective. The {+constricted glottis}
feature also explains why ejectives cannot themselves be voiced; the glottis is closed, which means it
cannot vibrate. In the remainder of this paper, the differences between the traditional sounds and
glottalic sounds is assumed to be:
-voice
+constricted glottis
+raised
In Italic, the series under discussion resulted in the reflexes /b/, /d/, /g/, /gu/, and /u/.
There is little confusion amongst these sounds; they are all well-attested, especially in Latin.
Typologically, these are all extremely common sounds among the world's languages. The following
examples of these reflexes and their reconstructions will help show how Italic is linked with PIE.
The labial sound (/b/ or /p’/) ended up as /b/ in Latin. An example of this is the word for
"strong": *bel- or *p’el- > Latin de-bilis. The alveolar (/d/ or /t’/) became /d/: *domos or *t’omos >
Latin domus, "house". And the velar (/g/ or /k’/) became /g/: *genos or *k'enos > Latin genus, "race".

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika:

Zgłoś jeśli naruszono regulamin