UNIVERSITY OF BENIN

FACULTY OF ARTS
DEPARTMENT OF LINGUISTIC STUDIES

ASSIGNMENT

ON

FEATURE HIERARCHY

BY:
OGBUCHUKWU PEACE CHIAMAKA ART1801561
LEONARD FAVOUR OZIOMA ART1801533
NDULUE OBIANUJUNWA GLORY ART1801541
JENEKA PEACE URIAMEH ART1801527
NWAEDOZIE ONYINYE EVELYN ART1801548
ADAMS AGHOGHO EVARISTER ART1801401
AIFUWA NOSAKHARE ART1801378
EHIOGUN VICTORIA ART1801474
JAMES ESTHER GIFT ART1801525
NWADIKE PAMELA OGOCHUKWU ART1801547
AIGBE DAVID OSAMUDIAMEN ART1801421
OKOTURO TORITSEMOTSE ELIJAH ART1801580
OSAWE BRIGHT ART1801607
IKUKAIWE ANTHONY NDUKA ART1701301

COURSE CODE: LST322 (NON-GENERATIVE PHONOLOGY)

COURSE LECTURER: DR. GLADYS IKHIMWIN

NOVEMBER, 2022.

1
 FEATURE HIERARCHY

Feature hierarchy is an offshoot of autosegmental phonology because it also makes use of milti-tiered

hierarchical structure, the CV skeletal and association lines which can act independently of one another

(Goldsmith 1995). Under feature hierarchy, segments are represented by features which are organized in

hierarchy of tiers. A segment is represented by a hierarchy of features which are organized into tree. This

hierarchy tries to express the dependent nature of the features as they act in groups to perform phonological

processes. Under this theory, segments are “geometrically organized features, each with some autonomy, but

also with some dominance relationship with other features”.

Mascaro (1985) and Mohanah (1983) are considered as the pioneers of feature hierarchy, stating that there is

a level of mutual dependence of articulators.

Clement (1985) called this feature geometry in his seminar paper titled “Geometry of Phonology Features”.

He made the theory to work on the weaknesses of the standard theory. He also asserts that one of the major

goals of feature geometry is to provide a theory of representation in which phonological processes can be

stated in a hierarchical manner. According to him, “the justification for the hierarchical feature tree is

premised entirely on the study of cross-linguistic generalizations concerning common types of phonological

and phonetic processes”. Since “it is well known that phonological processes may involve laryngeal features

without affecting supralaryngeal features”, examples including “rules of voicing assimilation, aspiration and

deaspiration”.

2
 Clement (1985) Feature Hierarchy tree is presented below:

X
|
Root

Laryngeal Supralaryngeal

Spread Constricted Voiced Manner Place

Nasal Continuant Consonantal Lateral Strident Sonorant

Round

Coronal High

Anterior Labial Back

Distributed

3
In the tree above, the terminal nodes occur at the base of the tree. On the top of the tree, we observed the
various classes in which the features are grouped.

Feature hierarchy formally encodes groups of features under nodes in a tree: features that commonly pattern

together are said to share a parent node, and operations on this set can be encoded as operation on the parent

node.

One node in feature geometries is the Laryngeal node. The Laryngeal node is an organizing node that

dominates the features of the larynx, usually taken to be [voice], [constricted glottis] and [spread glottis]. It is

common for these three features to pattern together in the phonology of world’s languages to the exclusion of

every other feature, and in feature geometry, this follows from the tree representation. Similarly, feature

geometries generally include a Place node and a Manner node which are dominated by the Supralaryngeal

node. The Place node is the dominant node of the place features, which also often pattern together.

The Root node is the topmost node of the feature tree and works as the formal organizing unit of the segment,

and in some frameworks encodes the major class feature such as [consonantal], [sonorant], and [approximant].

Some features such as [nasal] and lateral] are sometimes dependent on the root node, or sometimes of a

Supralaryngeal node along with Place. Other features such as [anterior] and distributed] are usually dependent

from the Coronal place feature.

Sagey (1986) proposed other important models. Models vary widely in the number of hierarchical nodes and

in how consonant and vowel features are treated. She based her feature hierarchy tree in the configuration of

the vocal tract.

4
 Sagey (1986) Feature Hierarchy tree is presented below:

Root

Laryngeal Continuant

Consonantal

Constricted Stiff vocal cords Supralaryngeal

Glottis Spread Slack
Glottis vocal cords
Soft palate Place

Labial Coronal Dorsal

Anterior Distributes

Back High Low

In Sagey’s feature tree, the laryngeal and supralaryngeal are part of the root node. In appending to the root

node are the other class nodes which are the continuant and consonantal. From Sagey (1986) model, ‘X’ is the

root. Laryngeal controls the activities of the glottis (voicing), while the supralaryngeal controls those above

the glottis.

Kentowicz (1994:452) introduced the major class feature in which the distinctive features are arranged as

proposed by Sagey (1986);

5
1. Major class features [syllabic], [consonantal] and [sonorant]. McCarthi (1988) and Hall (1992) said

that the [+consonantal] and [+sonorant] should form the root of the feature tree. This root node is the

only phonological motivated constituent in the hierarchy.

2. The place feature node which are [labial], [coronal] and [dorsal].

3. Tongue – body feature like [high], [low] and [back].

4. Manner feature: this includes the [continuant], [nasal] and [lateral].

Halle (1992) model, we take a look at only the features that are necessary.

Using the tree diagram, we can present the homorganic nasal [N] and [b]

/N/ /b/
[+cons, +Sonorant] [+cons, -sonorant]

[-cont] Supralaryngeal Supralaryngeal [-cont]

Soft palate Oral place Oral place

[Nasal] [Coronal] Labial

PROBLEMS OF STANDARD FEATURE THEORY

1. The standard feature theory focuses on the segmental properties of the features without reference of
their hierarchical organization. Thus, the features composing a sound are represented as an
unorganized bundle, allowing each feature to act independently of all others.

6
2. Certain features do not define any national class and thus, play supplementary role in the sub-
classification defined by another feature. For example, the features [anterior] and [distributed] only
seem relevant for coronal consonants but these feature values have also appeared for velar or glottal.
Although, by default, these consonants will be [-distributed], it will also allow the possibility of a
contrast [+/- distributed], velar or glottal, which is not correct.

3. Certain feature form recurrent groupings with regard to phonological rules (Clement 1985). For
instance, assimilation dealing with [high], [back] and [low] are common in vowels. However, it would
be rather unusual, if not impossible to witness an assimilatory event among [high], [nasal] and [ATR].
In other words, we should have a principled way to show that the feature [high], [back] and [low] have
a special affinity, whereas the other three, [high], [nasal] and [ATR] do not enjoy such relationship.

4. Another problem that existed in standard feature theory with respect to the implicit claim was that, if
one value of a feature denotes a natural class, then the other value of the same feature should do so
also. For example, considering the two features, [anterior] and [coronal], used to designate the places
of articulation, the following combinations are possible: [+anterior, -coronal]

Labials: [+anterior, +coronal]

Alveolars: [-anterior, +coronal]
Palatals: [-anterior, +coronal]
Retroflex: [-anterior, +coronal]
Velars: [-anterior, -coronal]
Uvulars: [-anterior, -coronal]
As a result, the following natural classes are obtained [+coronal]: (alveolars, palatals, retroflex).
While [-coronoal]: (labials, velaer, and uvulars).
This problem with such groupings is that, although the former, [coronal], is frequently attested in
phonological rules of languages, the latter, [-coronal] class is never found for a phonological event.

THE ACTIVE ARTICULATOR IN FEATURE HIERARCHY

In feature hierarchy, features are organized in some logical orders and these features are organized and
centered among six (6) active articulators;
1. The lips (labial articulator which executes the feature [round]),
2. The tongue blade (coronal articulator which executes the feature [anterior] and [distributed]),
3. The tongue body (dorsal articulator which executes the feature [back], [high] and [low]),
4. The tongue root (radical articulator, which executes the features [ATR] and [RTR]),
5. The soft palate (which executes the feature [nasal]), and
7
6. The glottis (glottal articulator which executes the features spread glottis, constricted glottis and the
feature voiced) as seen in the Sagey (1986) feature tree.

Considering an example for vowel elision:

/ada # obi/ /adobi/
Daughter # throne daughter of king

CV Tier: V C V V C V V C V V C V

a a o i a a o i

d b d b
[-syl] [-syl] [-syl] [-syl]
[+syl] [+syl] [+syl]
[+high] [-high] [+high]
[+syl]
[+high] [+syl] [+syl]
[-high] [+high]

In the above example, the second vowel “a” is not connected with any association line showing that is has
been elided.