Chapter 6
Naturalness and strength
Phonologists have used the term NATURALNESS to refer to the
fact that there is a phonetically well motivated relationship not
only between the allophones of a phoneme, but also between the
various phonological manifestations of a morpheme.
Naturalness can be approached in terms of MARKEDNESS.
In the sense that what is NATURAL is said to be UNMARKED, and
what is NOT natural is the MARKED.
Classes of sounds and individual segments both are affected by
the same phonological processes that tend to be made up of
segments which are phonetically natural.
The features combinations in a language that recur again and
again are said to be unmarked combinations (natural).
Examples of MARKED feature in English:
voiceless sonorants are much less common than their voiced
counterparts. Likewise, voiceless approximants are less common
than voiced ones. also, sounds produced with the velaric
airstream mechanism (i.e., clicks) and sounds produced with the
glottalic airstream mechanism (i.e., implosives and ejectives) arc
less common than sounds produced with the pulmonic airstream
mechanism. Front rounded vowels are rare but front unrounded
vowels are not, Nasalized vowels.
Languages typically have both natural segments and natural
phoneme inventories. Phonemic inventories tend to be
SYMMETRICAL. Also, observation can be restated in terms of
naturalness. As a rule, creating symmetrical phoneme Inventions
entails maximizing the use of a few phonological parameters
(economical features).
Natural Classes
Natural class: is a set of sounds in a language that share one or
more (phonetic) features, it is determined by participation in
shared phonological processes, described using the minimum
number of features necessary for descriptive adequacy.
Example of natural class:
The set containing the sounds /p/, /t/, and /k/ is a natural class of
(voiceless stops)
The set containing the sounds (/b/, /d/, and /g/) is a natural class
of (voiced stops)
(/f/, /θ/, /s/, /ʃ/, and /h) is a natural class of (voiceless fricatives)
Natural Segments
The segments which condition or undergo a phonological process
do share in each instance some phonetic characteristic.
Example about natural segments
1-Palatalization of velar consonants occurs in the context of front
vowels (especially high ones like [i]) which are themselves
produced with the tongue approximating the hard palate.
2-Labialization occurs in the neighborhood of labial vowels like [u]
which are themselves produced with rounded lips.
3-Nasalization of vowels occurs when they are adjacent to nasal
consonants.
4-Voiceless consonants may acquire some voicing when
juxtaposed with inherently voiced segments like vowels or
sonorants (such as nasals).
Naturalness and Markedness are not absolute concepts. Rather,
they are both relative. What is marked or unmarked will often
depend on the circumstances. For example, between two nasal
consonants, or before a nasal plus consonant cluster like [nd],
nasalized vowels would be unmarked. It would be somewhat
unusual for vowels occurring in those contexts to have no
nasalization. Using an oral [æ] in [mæn] is marked but using a
nasalized [i] and saying [min] is unmarked.
The same would apply to palatalization of velars before high front
vowels or labialization before rounded vowels. Markedness
cannot be interpreted with total disregard for context.
OBSTRUENT (or NONSONORANT): fricatives and stops (together
with affricates) form a natural class (These sounds share the
phonetic characteristics of having very significant obstruction in
the oral tract and of being typically voiceless, also tend to display
similar phonological behavior.) In English obstruents are more
heavily voiced at the beginning of a word than they are word-
finally. It is natural for obstruents to be voiceless. But, many
languages have more voiceless obstruents than voiced ones and
some have no voiced obstruent phonemes at all.
Phonological strength hierarchies
Beside the notion of assimilation and dissimilation that can be
discussed regarding natural phonological processes. There are
other processes for example, the notions of STRENGTHENING
(FORTITION) and WEAKENING (LENITION). These two concepts
are dependent on each other (they are relative).
Phonological strength hierarchy (> indicates a step towards a
'weaker' pronunciation):
(a) VOICELESS > VOICED SONORITY is the parameter
(b) STOP > AFFRICATE > FRICATIVE > APPROXIMANT > ZERO
manner of articulation is the parameter (Stops involve the
strongest obstruction and approximants the weakest)
Sonority hierarchy
1 voiceless obstruents (e.g. t s k) least
sonority
2 voiced obstruents (e.g. d g ß z)
3 nasals (e.g. m)
4 liquids (e.g. r)
5 glides (e.g. w)
6 vowels (e.g. a o) greatest sonority
The sonority hierarchy is an inverse restatement of the
strength hierarchy. On the face of it, the two
phenomena appear unrelated. However closer
examination reveals an interesting relationship,
especially where language evolution is concerned. To
understand this relationship, we need to draw a
distinction between SYNCHRONIC and DIACHRONIC
approaches to the study of language. Synchronic
linguistics studies the state of a language during one
period in its history. Diachronic (or historical) linguistics
on the other hand, studies the evolution of a language
during successive periods. Both synchronically and
diachronically, voiceless consonants change into voiced
ones in environments similar to those where the
reduction in the strength of the obstruction in the
production of consonants takes place.
One process may facilitate the other. It is significant
that the sounds at the weak end of the strength
hierarchy are typically voiced, while those at the strong
end are normally voiceless. Voiced sounds are weaker
than their voiceless counterparts — [d] is weaker than
[t], [zl is weaker than [s] and so on.
LABIAL > ALVEOLAR > VELAR This dimension of the
hierarchy seems to be valid for the languages of
Western Europe, like Danish, where in intervocalic
position, velar /g / undergoes the most extreme form of
lenition (i.e. weakening), being deleted altogether;
alveolar /d/ is moderately weakened, being changed to
the fricative [ծ] but labial /b/ remains unchanged.
The universal validity of a place of articulation hierarchy
is doubtful. In many Bantu languages, for instance, it is
the labial place of articulation that is weakest. Many
languages in this family historically weakened and
completely dropped the labial stops /p / or /b / while
retaining the alveolar and velar ones.
Proto-Bantu /p / is equally prone to lenition *p > h in
Sukuma (Tanzania), Rundi (Burundi), Pare (Tanzania)
etc.; *p > ɸ in Pokomo (Kenya), and Rimi (Tanzania).
Velars and alveolars do not undergo lenition to the
extent that labials do. Given this evidence, it would be
unwise to insist on the universality of a strength
hierarchy based on place of articulation.
VOICELESS STOP > VOICED STOP > VOICELESS
AFFRICATE > VOICED AFFRICATE > VOICELESS
FRICATIVE> VOICED FRICATIVE > NASAL > liquids>
glides>vowels
One type of consonant not included which there is
strong evidence for is the GEMINATE consonant like [t:]
or [d:], usually represented by doubling consonant
letters ([tt] or [dd]).
Gemination occurs when two identical consonants are
adjacent to each other in the same syllable as in English
penknife [pen:aif]; in other words, gemination occurs
when a particular segmental articulation is prolonged to
cover what would otherwise be two distinct segments.
The strength hierarchy, re-stated now in [6.15], is
manifested in synchronic phonological alternation and
in historical sound change in numerous languages:
GEMINATE VOICELESS STOP > GEMINATE VOICED
STOP > VOICELESS STOP > VOICED STOP > VOICELESS
FRICATIVE > VOICED FRICATIVE
The relative strength remains the same with other
manners of articulation. Synchronically, in many
languages in positions of weakening, for instance
between vowels, geminate segments like [t:] (spelled
with tt in the examples which follow) alternate with
plain segments like [t].
6.3 Explanations of naturalness
The assimilatory phonological processes of the kind
which basis in articulatory phonetics. We can point to a
good articulatory reason for a process like
palatalization, labialization or nasalization. Where
possible, adjacent sounds are made similar to each
other so that one avoids using any more effort than is
required to ensure that one is understood by the
addressee.
But while assimilation makes the task of speech
production easier, it can make speech perception more
difficult. It is easier to discriminate between sounds if
they are very different from each other than it is to
distinguish them when they are very alike.
To counterbalance assimilation, there are natural
processes which have the effect of enhancing
differences between sounds. These facilitate the task of
the hearer.
Voice dissimilation in Kirundi was introduced as an
example of dissimilation. When this rule applies, the
prefix and the root end up not having the same value
for voicing, which makes them more different from each
other than they would otherwise be.
Vowel patterns also frequently obey the principle of
MAXIMUM PERCEPTUAL DIFFERENTIATION. The set in
[6.16] turns up in language after language with a three-
vowel system
The choice of these vowels is not accidental. These three vowels
occupy the most peripheral positions in vowel space: [i] is the
highest front vowel, [u] is the highest back vowel and [a] is the
lowest vowel. Perceptually they are maximally distinct.
If a language has only three vowels, in order to avoid hearers
getting them confused (and as a consequence getting the
meanings of words in which they occur confused), it is almost
invariably the three vowels [i a u] that are selected. Languages as
diverse as Greenlandic Eskimo and Australian Pitta-Pitta have /i a
u/ as their only vowel phonemes.
Languages with a five-vowel phoneme system consisting of/i e a o
u/, if they have rules which neutralize vowel distinctions in certain
environments, tend to maintain the opposition between /i a u/ in
the places of neutralization. An example should clarify this. In
many Bantu languages with the five vowel phonemes /i e a o u/,
only the peripheral vowels /i a u/ occur in noun class prefixes
which mark the class (or gender) of a noun, as you can see in the
Luganda forms below:
