Kwartalnik Neofilologiczny
XXV, 3 (1978), p. 275-88
Warszawa
=======================================================
V. Y. Plotkin
University of Kishinev,
USSR
It is a widespread belief among linguists nowadays
that phonology has passed its days of glory and finds itself at a low ebb. In
my opinion, if it is to regain its vigour and achieve new heights, it must be
ready to reappraise much of the cherished legacy from its brilliant past. Among
other things it should be prepared to reconsider the pivotal place of the
phoneme in its theoretical edifice. A number of time-honoured problems, like
phonemic modelling, distinctiveness, neutralization etc. will then lose much,
if not all, of their former significance in phonological theory.
The crucial point in the development of phonemics
was reached when experimental investigations had demonstrated the impossibility
of accepting the phoneme as the basic unit in the production and perception of
oral speech. The inability of the phoneme to play this role is explained by the
simple truth that speech production and perception are cerebral activities
first and foremost, while the sound chain is the vehicle for their
externalization. The discovery and study of the phoneme as the language quantum
of audible speech sound was a great step forward in the history of linguistics.
It was quite natural for the observable phenomenon of sound to become an object
for theory before the internal cerebral mechanism of sound production and
perception. But this chronological precedence cannot justify the treatment of
the phoneme as the primary language unit on the plane of expression.
The preponderance of the oral and aural aspects of
speech over the underlying cerebral aspect in classical phonological theory is
an error that linguistics should waste no time in correcting. It is high time
for phonology to revive the pioneering ideas of J. Baudouin de Courtenay, who
did not make this mistake, whose psychologism, as F. Grucza has correctly pointed
out, “should not be treated as a step backwards in the development of
contemporary phonology” [1].
Just as the sound chain is made up of language
quanta designated as phonemes, so is the chain of cerebral activities leading
to the innervation of the mechanism for the production and perception of oral
speech segmentable into language quanta of a different level.
The need for the linguistic investigation of the
innervation stage of speech is obvious. As early as in the mid-50s R. Jacobson
and M. Halle expressed an interest in “the nerve impulses sent from the brain
to the effector organs”, which unfortunately were then “not yet open to a
precise analysis” [2]. “As to the transformation of speech components by the
nervous system, we can, for the time being, at best only hazard what
psychophysiologists have referred to as ‘a mere speculative assertion’ ” [3].
The following two decades apparently brought little progress in this field,
as “descriptions of speech and hearing in terms of innervations instead of articulations
seem to be lacking, interesting as such research should be” [4]. Phonemic
theory can enjoy the advantages of an impressive body of phonetic data
accumulated over more than a century, whereas the material on the innervation
quantum is scanty. But this is no reason why the notion should be neglected or
at best admitted as subsidiary to phonemic theory. The theoretician’s task is
not only to follow the experimentalist by interpreting his findings, but –
perhaps much more important – to lead the experimentalist by sketching areas of
scientific interest. Therefore, the lack of data on the nervous mechanism of
speech should not deter us from positing, on purely theoretical grounds, the
existence of linguistically relevant units embedded in this mechanism.
To designate the smallest unit of language, the
quantum of articulatory and auditory innervation, there is no better term than
J. Baudouin de Courtenay’s kinakeme, a blend of kinema, the
quantum of articulatory movement, and akousma, the corresponding
acoustic quantum [5].
The kinakeme covers practically the same ground as the notion of “distinctive feature”, but differs from the latter in implying the status of a language unit in its own right, not a mere tool for the identification and description of the phoneme as the basic unit of phonology. The status of a language unit presumes that kinakemes must be organized in paradigmatically structured systems and combine with one another syntagmatically to form units of higher ranks. Like the other two phonological units, the phoneme and the syllable, kinakemes may be used to carry semantically loaded language units.
The present paper demonstrates these properties of kinakemes in a number of languages, mostly European.
The paradigmatic
organization of kinakemes brings out very clearly the difference between the
notion of kinakeme and that of distinctive features. The latter are invariably
listed as items in inventories. R. Jakobson, M. Halle, N. Chomsky elaborated
universal catalogues of all the distinctive features known in human languages.
These elaborations have no doubt contributed greatly to the study of
anthropophonic possibilities and their actual implementation; however, as E.
Fischer-Jørgensen notes, items in a universal set cannot be units in the system
of an individual language [6]. In fact, N. Chomsky was quite explicit in
proclaiming the language-independent nature of his catalogue [7]. This approach
clearly reflects the treatment of distinctive features as ancillary instruments
in the description of the central phonological unit, i.e. the phoneme. It might
be argued that R. Jakobson did recognize the distinctive feature as the
ultimate unit of phonology and thereby rejected the pivotal role of the phoneme
[8], that N. Chomsky even discarded the very notion of phoneme. But as long as
one speaks of “features”, one cannot avoid ascribing them to some higher
entity, which is thereby assigned the central position in the theory,
regardless of pronouncements to the opposite effect. From this viewpoint N.
Chomsky’s replacement of the term “phoneme” by “sound” or “segment” does not
change anything in the subordinate status of their “features”. A
language-specific kinakemic system always consists of two subsystems, vocalic
and consonantal. This dichotomy results from the structural requirement for
contrast within the third unit in the phonological hierarchy, the syllable. But
the two subsystems are not rigidly separated, as they rely on very much the
same anthropophonic substance. It is not unusual therefore for languages to
utilize certain consonantal kinakemes within the vocalic crest of the syllable,
and vice versa, vocalic kinakemes may be found on its consonantal slopes:
well-known instances are consonantal kinakemes in syllabic sonants and vocalic
kinakemes in semivowels [9].
Another dichotomy divides all kinakemes into two categories, modal and locational. The dichotomy affects both subsystems. Modal kinakemes are concerned with the origin of sound and the vertical dimension of the vocal tract (the sonority axis in R. Jacobson’s terms); consonantal modal kinakemes determine the mode of obstruction and the acoustic type of sound – tone or noise, while their vocalic counterparts deal with the height of the vocal tract. Locational kinakemes, vocalic and consonantal, function on the horizontal plane (R. Jakobson’s tonality axis), activating certain areas along the vocal tract.
The two dichotomies described above seem to be
universal for all languages and affect both subsystems. Further dichotomies are
specifically vocalic or consonantal and display differences in individual
languages. In the following paragraphs we shall analyse consonantal kinakemic
subsystems in a number of languages.
The simplest among such subsystems are found in
Polynesian languages. In Maori each category is divided into two kinakeme
groupings: the modal kinakemes are obstructional or phonal, the locational
kinakemes are articulatoral or pointal. Each of the four groupings corresponds
to one of the four essential choices in the innervation of the speech organs
for the production of consonants, viz. the type of obstruction, the acoustic
nature of the sound, the active articulator involved and the point of
application for it. All the four choices are strictly binary in Maori, and each
of the four kinakeme groupings consists of two kinakemes; we shall designate
these groupings, intermediate between categories and kinakemes, as oppositions
[10]. The obstructional opposition contains the kinakemes of occlusion and constriction;
the phonal opposition is made up of the kinakemes of sonority (tone) and
discordance (noise); the articulatoral opposition embraces the kinakemes of
labiality and non-labiality (linguality); the pointal opposition consists of
the kinakemes of dentality and postdentality. In the locational category
articulators and points are described on a relative basis as central or
peripheral. The central articulator is the tongue or, in some languages, only
its front, while the lips are treated as precentral. The dental or alveolar
area usually serves as the central point of articulation.
Thus four tiers are available in the consonantal
kinakemic subsystem of Maori: the subsystem itself, two categories, four
oppositions with eight kinakemes [11]. These four tiers seem to be universal
for all kinakemic systems in all languages. Their binary organization may be
viewed with suspicion by many linguists, for binarism has been the object of
sharp (and often quite justified) criticism as something artificial with little or no
foundation in linguistic reality. In this connection a quotation might be in
point from a linguist who does not, generally speaking, favour binarism: “As to
binarism, phonemes cannot have binary ‘distinctive’ features because neither
our organs of speech nor our ear display a binary organization. As a model of
how the human mind works binarism is appropriate merely for the few who see the
world in black and white only. Articulation and perception may function in
steps, but that does not imply a binary model. Binarism may be a correct
representation of certain innervation
processes” [12]. On the whole, the role of binarism in language systems seems
to increase down the language hierarchy: while higher semantic systems display
little or no binarism, phonemic systems are noticeably binary, but less so than
kinakemic systems, which are lowest in linguistic rank.
We turn now to the Indo-European languages of modern Europe. In the English consonantal subsystem we find, of course, the tiers of categories (modal and locational), oppositions and kinakemes; but what serves in Maori as the substance of oppositions is utilized otherwise in English. The subsystem retains the grouping of modal kinakemes into obstructional and phonal, of locational kinakemes into articulatoral and pointal, but these groupings are not oppositions, since each of them contains four kinakemes instead of the two in Maori. For instance, the Maori obstructional opposition consists of the kinakemes of occlusion and constriction, whereas English has, besides these two, also the negative kinakemes of non-occlusion and non-constriction. That means that English has an opposition of the mutually contradictory kinakemes of occlusion and non-occlusion, which we shall briefly designate as the opposition of occlusion; similarly, there is an opposition of constriction. The tier of oppositions is thus quite different in substance from its Maori namesake.
The grouping of kinakemes into obstructional and
phonal, articulatoral and pointal does not belong to the tier of oppositions in
English, but forms another tier, intermediate between categories and
oppositions, which will be designated as subcategorial. The difference between
Maori and English can be described as coalescence of subcategories and oppositions
into one tier in the former language and their separation into two tiers in the
latter.
A subsystem with such coalescence is simple; a
subsystem is described as developed if it contains a separate tier of
subcategories. The test for the development of a subcategory is combinatory.
Two kinakemes that form an opposition, being mutually contradictory, cannot
enter the same phoneme. Thus in Maori the kinakeme of occlusion is incompatible
with that of constriction, but in English, where these two kinakemes belong to
two different oppositions, they are combinable, for instance, in affricates. In
the case of an undeveloped subcategory, which is an opposition at the same
time, all the phonemes are divided into no more than two classes with one or
the other kinakeme of this opposition. The development of a subcategory
provides conditions for the creation of three or four phoneme classes. A
phoneme may then contain not one, but two kinakemes of the subcategory in
question, one from each of the oppositions into which the latter is split. Four
patterns of kinakeme combination are possible: two patterns with a positive and
a negative kinakeme, a pattern with two positive kinakemes and a pattern with
two negative kinakemes. Usually only three of the patterns are realized in
phoneme classes.
Kinakemic subsystems may be developed fully or
partially. Full development implies the presence of all the four possible
subcategories with eight oppositions and sixteen kinakemes. English presents
such a picture of full development: besides the development of the
obstructional subcategory shown above, it has the oppositions of sonority and
discordance in the phonal subcategory, the oppositions of precentrality
(labiality) and postcentrality (faucality) in the articulatoral subcategory,
the oppositions of precentrality (dentality) and postcentrality
(palatovelarity) in the pointal subcategory [13].
A kinakemic subsystem is developed partially when
only some of its subcategories are developed, while the others remain
undeveloped.
The development of the obstructional subcategory
leads to the realization of three phoneme classes: occlusives, constrictives
and occlusive-constrictives. The occlusives naturally include the nasals /m n/
of all the languages under discussion and /ñ ŋ/ found in some of them. Nowhere is nasality represented as
a separate kinakeme, being the product of occlusion combined with the phonal
kinakemes of sonority and non-discordance in one phoneme. The opening of the
nasal passage is an inevitable result of the anthropophonic incompatibility of
a complete occlusion in the vocal tract with the achievement of sonority
without noise; the nasal innervation is not an independent action, but a
natural consequence of other actions. The constrictives include all fricatives
and some of the so-called liquids, e.g. English /w r j/, Polish
and Russian /l l'/, French /l r/.
The occlusive-constrictives, combining two positive obstructional kinakemes,
include, apart from all affricates, also certain liquids: /l/ in
some languages, /r/ in others. Both /l/ and /r/ may answer the
description of this phoneme class. The actual implementation of the combined
innervation, determined by the joint action of two positive kinakemes, depends
on the presence of other kinakemes in the phoneme. Affricates combine the two
obstructions successively at approximately the same location: in other words,
the obstructions are united in space, but not in time. For the reason stated
above this is impossible if the phonal kinakemes of sonority and
non-discordance enter the phoneme, and in this case the two obstructions are
combined otherwise. First, they may be realized simultaneously at two different
locations, united in time, but not in space, and that is known as the lateral
articulation of /l/. Secondly, occlusion and constriction may be quickly
interchanged, producing the vibrant /r/. Accordingly, both the lateral and the
vibrant sonants may theoretically be placed among the occlusive-constrictives;
but their actual kinakemic composition may differ from language to language.
Where no affricates are available the obstructional subcategory is undeveloped,
and neither /l/ nor /r/ is then occlusive-constrictive; this is the case in
French and Portuguese, in Dutch and the Scandinavian languages. Here /l/ and
/r/ are both constrictives with no kinakemic occlusion, and the difference
between these two phonemes with the same combination of modal kinakemes is
transferred to the locational category: /l/ retains the central point of
articulation, while /r/ is postcentral – uvular as in French or Dutch,
retroflex as in Swedish.
When the obstructional subcategory is developed and
affricates exist, either /l/ or /r/ joins the occlusive-constrictives; the
precise solution in the given language requires special investigation. The
markedly vibrant pronunciation of /r/ seems to indicate that it is treated as
occlusive-constrictive; this is the situation in the Slavic and Baltic
languages. The English and German /r/, on the contrary, are obviously
constrictives, therefore the /l/ of these languages must be described as
occlusive-constrictives.
Another problem for detailed investigation is the
status of semivowels. On the one hand, English /w j/ can be described in terms of consonantal kinakemes and
fit well into the same modal class of constrictive non-discordant sonants as
/r/. On the other hand, the French semivowels /j ў ŭ/ contain vocalic timbre
kinakemes, especially obvious in /ў/, which defies a consonantal description. Where /j/ is the only
semivowel, it permits very often both descriptions, consonantal and vocalic,
being in the latter case the non-syllabic variant of /l/. It is quite
possible that individual members of a language community differ as to the
kinakemic composition of /j/ in their idiolects, but the difference in the
innervation patterns is regularly eliminated at the effector level. That would
enable the hearer to analyse /j/ in his own way regardless of the way it is
produced by the speaker. Incidentally, this might shed light on the problem of
non-uniqueness in phonological solutions, phonemic and kinakemic.
The phonal subcategory, unlike the obstructional,
is developed in all European languages, and the kinakemes of its two
oppositions, sonority and discordance, may combine in phonemes in four
patterns: (a) discordance with non-sonority, (b) sonority with non-discordance,
(c) sonority with discordance, (d) non-sonority with non-discordance. However,
only Icelandic seems to have all the four phonal classes, exemplified by the fortes
/p t k/ as (a), the sonants /m n ŋ/ as (b),
their voiceless correlates as (c), the lenes /b d
g/ as (d). Elsewhere two arrangements are found with either (a b c) or
(a b d) as the patterns for kinakeme combination. Sonority coupled with
discordance is realized in voiced consonants like /b d g/, while the
discordant non-sonorants like /p t k/ are voiceless. The combination of the two
negative phonal kinakemes results in lenes, then the non-sonorant
discordants are fortes, usually aspirated to strengthen their noise
component. The latter pattern prevails in the Germanic languages, while the
Slavic, Baltic and Romance groups favour the former arrangement.
The development of the articulatoral subcategory
results in two oppositions of pre- and postcentrality. The positive kinakeme of
articulatoral precentrality enters the labial consonants. The positive kinakeme
of articulatoral postcentrality forms the faucal or the velar consonants,
depending on the kinakemic system of the given language: /h/ in English, /k g
x/ in Polish, Russian, Bulgarian, Romanian, Albanian, /k g
ŋ r x~h/ in German etc. The central
articulator, i.e. the tongue or its front, as the case may be, is activated by
the negative articulatoral kinakemes. Central consonants built in this way also
contain kinakemes of the pointal subcategory, which, when developed, embraces
the oppositions of pre- and postcentrality. The positive kinakeme of the former
produces predental or prealveolar consonants like English and Albanian /θ
đ/. The positive kinakeme of pointal postcentrality is available in many
languages, but its anthropophonic realizations vary widely. In English /k g
ŋ š ž
j/ there is variation from velar to palatal articulation depending on
the combination of obstructional kinakemes in the phoneme: occlusives are velar,
constrictives are palatal. A similar velar ~ palatal variation is observed in
French /k g ñ š ž r/ [14]. Pointal postcentrality is
realized palatally in phonemes like /š
ž tš dž/ found in many languages, also
/t' d' n' l'/ in
Czech, Slovak, Serbo-Croatian, Latvian, Albanian. The Spanish, Greek and
Icelandic /s/ seem to contain this kinakeme, and it appears to be emerging in
the Swedish retroflex /t d n
s l/.
The phonemes with two negative pointal kinakemes
form the central locational class of dental or alveolar consonants, depending
on the language. They include /t d n/ everywhere, /s z/ in most languages, except Spanish and Icelandic, where
this class includes /θ đ/ instead. The affricates
/ts dz/ where they exist are also
central.
The development of both locational subcategories
thus produces five locational series: (1) precentral articulatoral, (2)
precentral pointal, (3) central, (4) postcentral pointal, (5) postcentral
articulatoral. But only two languages among those under discussion, English and
Albanian, present such development. Most European languages display only four
series, designated above as (1), (3), (4), (5), due to the development of only
the articulatory subcategory. In French neither of the locational subcategories
is developed, and only three series are available, designated as (1), (3), (5).
We observe the following priorities in the
development of subcategories: in the modal category the phonal subcategory is
developed more easily, while in the locational category the articulatoral
subcategory takes precedence.
The evolution of new oppositions within developed
subcategories is impossible. But kinakemic systems can be further extended by
supplementary oppositions emerging within one of the categories. The
Indo-European languages of modern Europe possess supplementary oppositions of
two kinds. The languages of Eastern Europe – Polish, Lithuanian, Russian,
Belarusian, Ukrainian, Romanian, Bulgarian – have the locational opposition of
palatalization. The other supplementary opposition belongs to the modal
category – it is the opposition of protensity (length, gemination). In ancient
times it was spread over the entire Germanic region, where today it is
preserved only on the Scandinavian peninsula, in Swedish and Norwegian [15].
In the Romance group it is found in Italian. Supplementary oppositions are
instrumental in increasing, in fact almost doubling the phoneme inventory. It
is noteworthy that systems of writing distinguish supplementary oppositions
from the basic oppositions by using diacritics, not separate letters for
phonemes with positive supplementary kinakemes.
Like all other language units, kinakemes form
syntagmatic combinations. Many linguists equate syntagmatic grouping with
linear arrangement of segments. According to É. Benveniste, only segmentable
units can form syntagmatic classes; kinakemes (merisms in his terminology),
being non-segmentable, are denied this ability, although it is admitted that
phonemes are produced by the combination of merisms [16]. But is such
combination of units not a syntagmatic operation?
It must be accepted that
linearity is not universal on the syntagmatic axis. It is a common syntagmatic
property of language units beginning with the syllable, the smallest quantum of
articulate, segmentable speech. Phonemes are also arranged in linear
succession, but with much overlapping, which is a deviation from strict
linearity and an obstacle to phoneme segmentation of the speech chain. Phoneme combination, basically
linear, thus displays a good measure of simultaneity. As for intraphonemic combination of kinakemes,
simultaneity is its dominant principle.
The reign of simultaneity on the lowest linguistic
level is quite natural. As we go down the hierarchy of language units, their
numbers in paradigmatic systems decrease: there are fewer morphemes than words,
fewer phonemes than morphemes or syllables, fewer kinakemes than phonemes. But
the smaller the unit, the more numerous it becomes in syntagmatic lines: a
medium-sized sentence of three or four words might contain about a dozen
morphemes, several dozen phonemes, while the number of kinakemes will run into
hundreds. A linear arrangement of the smallest and therefore most numerous
units would lengthen the utterance beyond human abilities. Only in special
circumstances, where limiting the length of the message is not so vital, may
human communication become linear throughout, as illustrated by the Morse code
in telegraphy. The simultaneity of kinakemes is also inevitable for
anthropophonic reasons, since only the joint action of several kinakemes is
externalized in sound.
Kinakeme combinations are intraphonemic and
interphonemic. Laws of intraphonemic kinakeme combination are in part universal
or near-universal, in part language-specific. A universal requirement
determines a balanced numerical relationship between positive and negative
kinakemes within a phoneme. Positive modal kinakemes combine more easily with
one another than positive locational kinakemes, but limitations exist for both
categories.
Consonantal subsystems in the languages under
discussion possess three modal oppositions when the obstructional subcategory
is undeveloped, four when it is developed. Accordingly, a consonant phoneme
might, theoretically, contain three or four positive kinakemes or the same
number of their negative counterparts. In reality, however, both extremes are
avoided. German and French illustrate the ways in which this is done.
In French the
obstructional subcategory is undeveloped, and there is one opposition of
occlusion versus constriction, with the latter presumably its positive member.
A consonant has only one obstructional kinakeme, which is negative in occlusive
consonants like /p t k/. But in the other modal subcategory
the combination of two negative phonal kinakemes, i.e. non-sonority with
non-discordance, is prohibited. As a result, the joint occurrence of three
negative modal kinakemes is ruled out, and a consonant contains at least one
positive modal kinakeme as in /p
t k m
n ñ/, at most three as in
/v z ž/.
The German obstructional category is developed, so
there are four positive modal kinakemes. A consonant contains at least one positive
obstructional kinakeme and may contain both of them. In the other modal
subcategory, in contrast to French, the prohibited combination is that of two
positive phonal kinakemes, i.e. sonority with discordance, while their negative
counterparts are mutually combinable, as in /b d g/. Therefore
a German consonant cannot have all the four positive modal kinakemes, whose
possible number in a phoneme is from one, as in /b d g v
z/, to three, as in /pf
ts tš/. Thus German and
French, so different in their kinakemic and phonemic systems, display the same
range for the number of positive kinakemes in their consonants.
In Russian each of the modal subcategories
prohibits the combination of its two negative kinakemes and permits the
combination of both positive kinakemes. The minimal number of positive modal
kinakemes in a consonant is therefore two, but the maximal number is not four,
but three, because voiced affricates, which would present a combination of all
the four positive modal kinakemes, are absent from the Russian phonemic system.
Positive locational kinakemes do not easily
accumulate in consonants. There is usually one such kinakeme in a consonant,
either articulatoral or pointal. A combination of two positive articulatoral
kinakemes is present in labiovelar consonants wherever they exist as
monophonemes. A consonant with a positive articulatoral kinakeme usually
contains no pointal kinakemes at all [17]. All the locational kinakemes are
negative in consonants with the central articulator applied to the central
point. Supplementary locational kinakemes have wide combinatory capabilities.
Intercategorial combination of modal and locational
kinakemes in a consonant is usually subjected to language-specific limitations.
The dominant trend is to bar the inclusion of locational kinakemes into
phonemes with many positive modal kinakemes. This affects primarily the
occlusive-constrictives, which in some languages (English, Spanish) have no
locational kinakemes at all, in many others (the Slavic and Baltic languages,
Italian, Romanian, Albanian) admit only the kinakemes of pointal postcentrality
in /ts tš/; German alone has
affricates with both articulatoral and pointal kinakemes.
The product of kinakeme combination is the phoneme.
In the light of kinakemic theory phonemes appear as conventional, standard
blocks of kinakemes. The phonemic paradigm of a given language is shaped by
general laws and specific rules of kinakeme combination.
Some potential
combinations of kinakemes, not prohibited by the laws and rules, may fail to
become standard in a given language. This results either in complete gaps in
the phonemic paradigm, or in the phenomenon of “peripheral” phonemes, whose
integration into the paradigm appears to be incomplete [18]. An instance are the
combinations of the positive obstructional kinakemes of occlusion or
constriction and the positive phonal kinakemes of sonority and discordance with
the positive locational kinakeme of articulatoral postcentrality (= velarity).
In most Slavic languages only one of these two combinations is standardized
into a phoneme: this is the one with occlusion in Polish, Northern and literary
Russian, Bulgarian, Serbo-Croatian /g/, the one with constriction in Southern
Russian, Ukrainian, Belarusian /γ/. The non-standard kinakeme combination is not necessarily absent from
the language entirely, it might crop up in borrowings or expressive words and
even reach eventual standardization. This is the way in which the Czech
language acquired the phoneme /g/, still restricted etymologically, but
undoubtedly integrated in the paradigm. A precarious existence characterizes
the Ukrainian /g/, as in ганок ‘porch’, which is replaced by /γ/
[19], the
literary Russian /γ/, regular only in the
interjections ага, ого, угу, эге between identical vowels, and
the Polish /γ/ replaced by /x/, as in herbata
‘tea’.
The precise status of non-standard kinakeme combinations
is a moot point in phonemic theory. Lively discussions have taken place about
the consequences of gaps in phonemic paradigms for phonemic distinctiveness.
For instance, doubts have been expressed as to the distinctive relevance of
voicelessness in /x/ where no /γ/ exists. The
very concept of phonemic distinctiveness, however, is rightly rejected by F.
Grucza [20]. It has no place in kinakemic theory. What matters is the existence
of certain kinakemes in the language and the realization of their combinatory
potential. The failure of some kinakeme combinations to materialize or to
become standardized cannot alter the status of standard combinations in any
way; it may only influence the diachronic destinies of the latter by opening
ways for their change.
Interphonemic combination of kinakemes is governed
by language-specific rules, traditionally described as rules for positional
neutralization. Kinakemic formulations are also possible and seem to be
preferable, allowing greater generalization. Devoicing of Russian consonants in
word-final position, their voicing and devoicing in clusters are treated as
typical cases of neutralization. In kinakemic terms we state that the positive
phonal kinakeme of sonority can join the other positive phonal kinakeme of
discordance in the same phoneme only when a phoneme follows which does not
contain the latter kinakeme; the kinakeme of sonority, when permitted by this
rule, is imposed on the entire cluster of discordants.
Phonological
units are unilateral and have no meaning of their own. They provide shapes for
bilateral, semantically loaded units, usually in concatenation with one
another. But concatenation is not obligatory, and shapes for semantic loading
may consist of single syllables or phonemes. The kinakeme is not an exception
in this respect. Syllables serve as shapes for root morphemes and words,
notional or formal; non-syllabic phonemes can provide shapes for derivational
morphemes, sometimes formal like the Polish prepositions w, z,
the Russian в, с, к. Kinakemes may be used singly
as shapes for derivational morphemes.This is demonstrated by Russian pairs like
рваный ‘torn’ – рвань n. ‘rags’, мазать ‘to anoint’ – мазь n. ‘ointment’, where the
kinakeme of palatalization embodies the noun-forming suffix [21]. The kinakeme
of palatalization is widely used in Romanian nouns and verbs as a shape for
their inflexions: cf. muncitori ‘workers’, elevi ‘pupils’, (tu)
dormi ‘(thou) sleepest’. Syntagmatically this kinakeme enters the last
consonant of the stem. It follows that morphemic boundaries, contrary to the
traditional view, can run through a phoneme. The traditional rule remains valid
only for heterogeneous sound sequences – diphthongs and affricates. It must be
reformulated in the sense that morphemic boundaries cannot run between
successive parts of a phoneme, but are possible between its simultaneous parts.
Not only single kinakemes, but also non-phonemic
combinations of kinakemes are sometimes employed as morpheme shapes. Such
combinations may be smaller than a phoneme. The English inflexions spelt -(e)s
and -ed contain the same set of five negative kinakemes: the four
negative locational kinakemes and the phonal kinakeme of non-sonority; the two
inflexions differ in their obstructional kinakemes, the former being
constructive non-occlusive, the latter occlusive non-constrictive. Both
inflexions have shapes formed by the combination of seven kinakemes. The eighth
kinakeme found in the corresponding consonants /s z/ and /t d/
belongs to the opposition of discordance; determined by the last phoneme of the
stem, it remains outside the inflexion shape. Combinations of kinakemes larger
than a phoneme can also form inflexion shapes. Instances are found in Polish
and Russian, where, for example, the dat. sg. fern. inflexions in mapie, карте ‘to the map’ contain, besides the phoneme
/e/, also the kinakeme of palatalization for the final consonant of the stem.
Kinakemes are widely used as morpheme shapes in the Celtic languages [22].
In summing up we can state that the kinakeme possesses the paradigmatic, syntagmatic and semantic properties which are characteristic of other phonological units, and it fully deserves to be treated as one of them, the smallest item in the hierarchy of language units.
References
1. F. Grucza. (Review of:) F. H. H. Kortlandt. Modelling the Phoneme.
“Phonetica”, 30 (1974), 1, p. 60.
2.
R. Jakobson, M. Halle. Phonology and
Phonetics. In: R. Jakobson, Selected Writings, I. The Hague:
Mouton, 1962, p. 487.
3.
Ibid.
4.
H. Bluhme. Segmental Phonemes versus
Distinctive Features in English. Linguistics, 126, 1974, p. 17.
5. Cf. V. Y. Plotkin. Systems of
Ultimate Phonological Units. Phonetica, 33 (1976), 2, p. 82-83.
6. E. Fischer-Jørgensen. Trends in Phonological Theory: A Historical
Introduction. Copenhagen: Akademisk Forlag, 1975, p. 152.
7. N. Chomsky, M. Halle. The Sound Pattern of English. New York: Harper
& Row, 1968, p. 335. N. Chomsky. Language and Mind. New York:
Harcourt, 1972, p. 74.
8.
R. Jakobson. On the
Identification of Phonemic Entities. In his Selected Writings, 1.
The Hague: Mouton, 1962, p. 425.
9.
R. Jakobson, M. Halle. Op. cit., p.
478.
10. This veteran among phonological terms is given here an entirely
different meaning, denoting the smallest paradigmatic group of two kinakemes,
not a relation between phonemes.
11.
Cf. V. Y. Plotkin. Op.
cit., p. 84-86, 90.
12.
H. Bluhme. Op. cit., p.
17.
13. V. Y. Plotkin. Op. cit., p. 86-90.
14. Cf. R. Jakobson, J. Lotz. Notes on the French Phonemic Pattern. In
R. Jakobson, Selected Writings, I. The Hague: Mouton, 1962, p. 428.
15. V. Y. Plotkin. Is Length Phonologically Relevant in Icelandic?
Phonetica, 30 (1974), 1, p. 35.
16. É. Benveniste. Problèmes de linguistique générale. Paris:
Gallimard, 1966, p. 121.
17. A phoneme may contain one of the two kinakemes in an opposition or none
of them. In a kinakemic matrix of phonemes this is reflected by the signs of
plus, minus or zero respectively.
18. Cf. J. Vachek. On Peripheral Phonemes. Proceedings of the 5th
International Congress of Phonetic Sciences. Basel: Karger, 1965, p. 561.
19. В. С. Перебийнiс. Кiлькiснi та якiснi характеристики системи фонем сучасної украiнської лiтературної мови. Київ: Наукова думка, 1970, стор. 45.
20. F. Grucza. Sprachliche Diakrise im Bereich der Ausdrucksebene des
Deutschen. Beiträge zur allgemeine Sprachtheorie. Poznań 1970, S. 106
ff.
21. Т. В. Булыгина. О
некоторых аналогиях в соотношении семантических и звуковых единиц. Вопросы
языкознания, 1967, № 5, с. 82. Н. Д. Арутюнова. О значимых единицах языка. Исследования
по общей теории грамматики. Москва: Наука, 1968, с. 87-88. В. С. Перебийнiс. Op. cit, стор. 18-19.
22. Л. Г. Герценберг.
Морфологическая структура слова в ирландском языке. Морфологическая
структура слова в индоевропейских языках. Москва: Наука, 1970, с. 79-87.