The phonological view reveals an underlying structure of the position of the vowels to consonants. Vowels typically occupy a central position, whilst consonants are typically located marginally. By looking at combinations of phonemes, called phonotactics we can ascribe how certain combinations of phonemes are legal, whilst others, though plausible from an articulatory perspective are deemed illegal. We will look at these ‘rules’ and see how combinatory phonemes are binded together to form syllables.
2.1 English Syllable Structure
Generally speaking, English never has more than two vowels in a sequence in any one word, but consonants may appear in clusters of two, three, or more, as in the words ‘describe’, ‘crunch’ and ‘split’ (Kreidler, 1989:77). Such generalisations provide rules for the limitations, or constraints in which vowels and consonants can coexist (Roach, 2000). The English syllable minimally consists of one vowel (e.g. a, eye, ow!). Roach gives the example of one syllable words: `are’ /aː/, `or` /ɔː/, which are preceeded and followed by silence, he also includes isolated sounds such as m, used to communicate agreement, or ʃ, to ask for silence, as syllables. Some syllables have an onset which generally occur on the margins of vowels which are represented as the peak or nucleus of a syllable. Syllables with no onset are called a coda. In big the onset is /b/, the peak is /i/, and the coda is /g/; in free the onset is /fr/, the peak is /i/ and there is no coda; the word ox has a zero onset and the peak is /ɒ/, and the coda is /ks/ (Kreidler, 1989). The syllable structure can be illustrated in the following way:
S S S
O P C O P C O P C
/ b i g fr ii φ φ ɒ ks /
(Kreidler, 1989:75 )
Every language has a common pattern or code in which phonemes are arranged to form syllables, so that in English words can be realized by the formula (CCC)V(CCCC). By using these possible phoneme combinations, the study of phonotactics enables possible legal combinations which can occur in any language. Furthermore, a more refined analysis of phonology has divided the syllable into an onset and a rhyme, in which the peak (usually the vowel) and a coda (which is not always obligatory as in the word ‘me’) are known as a rhyme, giving the following appearance:
Syllable
Rhyme
onset peak coda
(Roach, 2004: 77)
Despite the theoretical CVC form Kreidler (1989) states that it would be feasible to expect all consonants as occurring in initial and final position and all vowels in between. But he maintains that although the CVC canonical form is common there are limitations on the consonants in initial and final position, so for example before final /ng/ or /nk/ there are only four possible vowels:
sing sang song sung
pink rank honk dunk
(Kreidler, 1989: 132)
These constraints argue Kreidler, not only affect spelling but also affect the way people pronounce and perceive vowels so that the vowel of ‘sing’ may be seen as the same, or different from the vowel of ‘seat’ depending on different speakers interpretation. It seems that although linguists have come closer to establishing agreement of an invisible line between Jesperson’s syllable valleys, there are still aspects which remain fuzzy as will be seen in the next section which deals with more ambiguous divisions of syllables.
2.2 Syllable division
As we have seen, the concept of tiers, as in CV phonology allows complex segments to be captured, highlighting divisions between clusters of consonants and vowels (Clark and Yallop, 1995). But despite this it does not provide a definitive explanation for deciding the division between syllables (Roach, 2004). Deciding where two syllables should be divided is far from straightforward, as demonstrated by the word `extra` (ibid.):
- e.kstra
- ek.stra
- eks.tra
- ekst.ra
- exkstr.a
Roach questions how such a division can be decided. He offers the theory of maximum onset principle which states the following problem: if a word like `athlete` has an agreed two syllables, where does the syllable boundary occur? The maximum onset principle works on the basis that as many consonants should be syllabified with a following vowel, providing that the resulting sequence is phonotactically legal. For the word `athlete` the decision as to whether /θl/ belongs with the first or second syllable is the phoneticians dilemma. Using the maximum onset principle the question to be asked is whether there are any words in English which begin with /l/? As there are the /l/ should be assigned to the second syllable. The second question is whether there are any syllables that begin with /θl/? As there are not the syllable boundary should be placed after /θl/ (ibid.). It can be established by this that the word `athlete` has two syllables /æθ/ and /liːt/. (Internet 1). Roach however, questions how words like `better` /betа/ can be divided. As the maximum onset principle tells us that the /t/ should be placed on the right-hand syllable, giving be.ta, but be, is not allowed according to the maximum onset principle. According to Roach, `the maximum onset principle must therefore be modified to allow a consonant to be assigned to the left syllable if that prevents a short vowel from occurring at the end of a syllable’ (ibid. p.75). He suggests the word bet.a as the most satisfactory. He also extends one more possibility when a consonant stands between a vowel and a consonant, as in `better` or `carry`, the idea that we could assign it to both syllables; for a consonant can be ‘ambisyllabic’ (ibid. p.78).
3 Constraints
Having explored what is possible we will now inquire what is not, and ask a question: why are there constraints on syllables? As we have seen, syllables can begin with various vowel/consonant sequences, however there are constraints on various sequences, for example, if we examine consonant sequences we can see that the combinations zb, zt, tl or dl are not possible, although three letter consonants such as spr, str, as in ‘spring’ and ‘string’ are (Celce-Murcia et al, 1996; Knowles, 1987). In the same manner no word ends with the sequence æh. Such constraints are known as phonotactic or phoneme sequence constraints, and limit the number of syllables which can be compounded in any language. All languages have such constraints based on the limits of the talker’s ability to pronounce sequences of sounds as one syllable and he phonotactic constraints provide a clue as to the number of syllables in a word (Internet 2: Harrington & Cox). To some extent the constraints on English syllable clusters relate to what is easy, so that a cluster such as /mk/ or /lpr/, not possible in English is difficult to produce even for a trained phonetician. There are also some clusters which used to be legal in English such as hnutu, and in some cases they are reflected in English spelling such as /wr/ as in ‘write’ and ‘wrong’ (Knowles, 1987). Whilst the clusters tl or thl are not difficult to pronounce, at some time in the past such clusters `ceased to be legal` (ibid. p.70).
4 Stress
English, like Germanic languages has stress timing. The difference between stressed and unstressed syllables is greater in English than in most other languages (Clark & Yallop, 1990; Celce-Mercia et al, 1996), it is also an extremely important characteristic of the syllable. In stress languages like English, a certain syllable within a word is perceived as dominant and it is this syllable which is said to carry the primary stress (Tsujimura, 1996). Stress can fall on any part of the syllable depending on the origin of the word and according to lexical status i.e. abstract ‘æbstræct (adjective) and æb’strækt (verb) (Roach, 2004) and such lack of predictability causes confusion for second language learners (Celce-Mercia et al, 1996). Perhaps for this reason the phenomenon of stress in English has received substantial attention for identifying its salience within syllables and words.
4.1 Identifying Stress in Syllables (for learners)
What are the characteristics of stressed syllables and what rules, if any, can we apply to syllables for learners? Unlike French or Polish where stress is received on the last and penultimate syllable respectively (Gimson 1964; Roach 2004), English stress placement is a matter much more complex, with examples of stress as independent of syllables whilst in some cases clearly being identified within syllable. We will look briefly at these two areas.
First, Katamba (1993) in his discussion of stress, uses evidence to show that stress is a prosody, i.e. not residing in any consonant or vowel, but rather ‘stress can hop from syllable to syllable’ (p.155) so that a word such as `democrat becomes demo`cratic. The causative factor being the suffix rather than the syllable. Katamba succintly concludes:
‘If stress were an integral part of the segments in a syllable it would not enjoy such mobility and independence.`
(Katamba, 1983: 155)
Therefore lexical status plays a greater role in stress placement as demonstrated with the word ‘abstract’, both a noun and an adjective: ‘æbstræct (adjective) and æb’strækt (verb) (Roach, 2004). So if stress does not reside in the syllable what rules can be given to a learner? Kreidler (1989) argues that there is a large measure of predictability about English stress and provides a definition of an important characteristic of stress syllables:
`The stressed vowel of a polysyllabic word can always be identified with the vowel of some monosyllabic word` (p.80).
Consequently the stressed vowel of `aroma` is the same as the vowel of `row`. This may be of help for learners looking for a stable rule to apply. Roach adds four more:
-
Whether the word is morphologically simple, or whether it is complex as a result either of containing one or more affixes.
- What the grammatical category of a word is (noun, adjective, etc.).
- How many syllables the word has.
- What the phonological structure of those syllables is.
(Roach, 2004:97)
Clark & Yallop (1990) also illustrate that the patterning of `word-stress` or `lexical-stress` in spoken English is not always based on word, so that phrases such as `a party` or `leave it` will have the pattern of single words with one short syllable, and similarly there is no difference between single words such as `array` or `arise` and two-word combinations such as `a ray` and ` a rise`. For Clark & Yallop (1990) this serves to illustrate that certain words (grammatical or orthographic) are unstressed, such as: `a`, `the`, `to`, accordingly they stipulate that stress can also be determined by syllable rather than meaning. However, despite the debate over the complexity of syllable stress, speakers can, when they come across an unfamiliar word, pronounce it with the correct stress (Roach, 2004). So that an imaginary English word like bago /bаgo/ would always receive stress on the first syllable.
To summarise, a definition of an English syllable depends on whether it is construed phonetically or phonologically. Phonetically it can be represented as containing an onset, a peak and a coda, phonologically it can be construed as something containing a vowel, with or without surrounding consonants. The classification of an English syllable may be abbreviated in the formula (CCC)V(CCCC) or by saying that the vowel (V) or the syllable may be `proceeded by up to three consonants (C) and followed by up to four` ((O` Connor (1973:200). Having addressed differences between phonology and phonetics and we can now apply some of these distinctions to the study of the Japanese syllable.
5 Japanese 日本語
Most apparent to learners of Japanese is that it is a di-syllabic language, meaning that vowels and consonants do not exist on their own as in an alphabetic script such as the roman alphabet; instead they exist in combinations such as /ka/, /ki/, /ku/, /ke/, ko/ and so on. The Japanese syllabary has five vowels and a simple set of thirteen consonants and two semi-vowels, all of which make up 104 syllables (Tsujimura, 1996), with each combination standing for one syllable. It is these syllables that are represented by each of the forty-six basic hiragana and katakana syllabaries (see appendix 1). The remaining syllables can be represented by the hiragana/katakana with slight modifications. It has a basic CV syllable structure:
(C1)(C2)V1(V2)(C3)
and can be classified as a tone language (Tsujimura, 1996; Shibatani, 1990; Kindaichi, 1978), but unlike its tonal cousin Chinese in which it is usually necessary for each syllable to be specified, in Japanese, the pitch of each syllable can be predicted (Shibatani, 1990). Furthermore unlike English which has syllables to mark word/phoneme boundaries, the accentual system of Japanese involves two units of counting; syllables and mora. We will first differentiate between mora and syllable before discussing timing and stress characteristics of the Japanese syllable.
6 Mora: Differences between Syllable and Mora
In order to contrast syllables within English and Japanese, qualification is first required between the term syllable and mora (Shibatani, 1990). An example of this can be demonstrated by studying the pronunciation of the word ‘London’. If a native speaker of English were asked to count the number of syllables occurring in the word they would probably answer `two’, but if a native Japanese were asked the same question the answer would be `four’ (Tsujimura, 1996). What is the reason for this phenomenon? In Japanese the word `London` written in katakana is ロンドン and is realised as four segments: /lo.n.do.n/, quite different from our earlier description of a syllable. The mora then is qualitatively different than the English syllable which remains problematic when attempting to divide certain words. On the contrary, the Japanese syllable has been described as `exceedingly simple` (Kindaichi, 1957:109) in structure, with each katakana representing a separate mora. We can further examine this notion of simplicity by at further examples. The Japanese word for autumn ‘aki` /a.ki/ [あき], has two morae: /a/ [あ] and /ki/ [き]. So far this resembles the English syllable, but if another consonant is inserted, giving the word for `instrument`, `gakki` /gakki:/, it becomes three morae, as does the word for ‘phone’ /denwa/: both are illustrated below:
/ga.k.ki/ [がっき] /de.n.wa / [でんわ]
Unlike speakers of English who divide words based on onset and rime, Japanese speakers divide words into smaller parts on the basis of mora or kana syllable. A mora then is `a unit that can be represented by one letter of kana and functions as a rhythmic unit in the composition of Japanese poems, e.g. waka and haiku (Shibatani, 1990). The mora can be realized by the following description:
1. (CC)V
2. The first part of a long consonant
3. Syllable final, or `moraic`, nasal /n/.
(Tsujimura, 1996:78)
The nasal /n/ makes reference to a /n/ when not followed by an accompanying vowel as in nendo /nendo/ [ネンド] (clay) or in the final /n/ of /mikan/ [みかん] for `orange`. These two words can be subdivided using mora as:
/ne.n.do/ /mi.ka.n/
Following an investigation into mora and syllable metrics, Gilbert and Yoneaka (2000) also show the difference between the English syllable and the Japanese syllable. The Japanese word pan /pan/ パン (meaning bread) contains two mora as onset-nucleus /pa/ plus the coda /n/: p.an. This compares with the English syllable which has an onset, nucleus and coda and is the length of the Japanese syllable. Viewed phonetically it can be presented in the following way:
(Gilbert and Yoneaka, 2000)
6.1 Syllable Debate
Should we dismiss the notion of comparison between the syllable of the mora? Despite the acceptance of the mora as separate from the syllable, Tsujimura (1996) showed that certain dialects would not divide a word like mikan into three parts, but two: mi.kan; therefore she contends that the `notion of syllable structure is not always missing in Japanese’ (p.80). Shibatani (1990) concurs with Tsujimura in distinguishing two types of dialects and hypothesizes that once Japanese was a ‘syllabeme language from which the more contemporary mora dialect developed’ (p.160). Shibatani therefore asserts that both mora and syllables play an important role in the Japanese language.
7 Mora Timing
Japanese is a syllable-timed language and has been variously defined as having a `monotonous` (Kindaichi; 1978), `machine gun rhythm` (Morio, 1997) with each syllable sounding the same in length; creating a distinctive and idiosyncratic sound peculiar to Japanese. This is promoted by the mora, which has a timed unit bearing the same length, so that if two sets of three-mora words are compared the duration of the two words are the same (Tsujimura 1996, Kindaichi; 1978). Unlike problems which face phoneticians with English words like ‘extra’ (seen earlier), Japanese speakers can distinguish syllables with relative ease (Kindaichi, 1978), for example, the word ‘Hiroshima’ pronounced slowly gives us hi-ro-shi-ma, with each mora given the same length. Within this, Japanese pitch creates unity by denoting where a word should begin or end. (Kindaichi, 1987:123). The equal length of mora syllables in Japanese poetry which stipulate seven-and-five syllable and five-and seven syllable meter have produced the phenomenon known as waka – a poem composed of 31 syllables.
In terms of English syllable structure similar haiku poems are possible, e.g:
In a potato
Those groans whose forced prayers change nought
Can never occur
(Gilbert and Yoneaka, 2000)
Although it has the 7-5-7 timing `to the eye and ear, to the sense of counting it has no meaning (Blyth, 1963). For Gilbert and Yoneaka this further illustrates the difference between a syllable and a mora, the former consisting of an onset, nucleus and coda, the latter as a set of phonemes (C)V. To sum up, the Japanese syllable is typically short and composed of (C )V, with a certain metrical pattern which eludes the stress-timing of the English syllable.
8 Accentuation
If a textbook definition were applied to Japanese, Japanese dialects would be called tone languages Shibatani (1990). An important difference between English and Japanese syllables lies within how stress and pitch are realised on words (Tsujimura, 1996). Unlike English where syllables receive prominence, Japanese is a pitch-accent language. The location of the pitch must be learnt with the word as part of its pronunciation, therefore the accent is ‘lexically indicated’ (ibid. p.74). The following words serve as an example:
1. sora ‘sky’ (high-low) 2. kawa ‘river (low high)
(Tsujimura, 1996:74)
Like English any given word can, without changing the vowels and consonants, be made into more than one distinguishable word such as the word ‘hashi’ [はし] which has three possible meanings: ‘bridge’, ‘edge’ and ‘chopsticks’, depending on whether the pitch is high or low:
1. hashi ‘chopstick’ (low-high) 2. hashi ‘bridge’ (high-low) 3. hashi ‘edge’ (high-high)
(Shibatani, 1990:179)
But unlike English the stress comes from the variation in pitch. In English, one factor alone such as high pitch is not sufficient to indicate stress.
9 Constraints
For Japanese, syllables can end on a vowel and can start with a maximum of one consonant, but no clusters are permitted. In Japanese a syllable can also end with an /n/ [ん], or if the syllable is not at the end of a word, with the consonant that begins at the next syllable (providing the consonant is /p/, /t/, /k/, /s/, or /c/). The group of words below include what is phonotactically legal and illegal in Japanese:
■ legal: /e/, /se/, /te/, /kantan/, /natoo/, /ken/
■ illegal: /nat/, /mak/, /bum/, /ste/.
The constraints operating in Japanese and English then are very different (Internet 4: Gasser, 2002).
English and Japanese Syllables: Table of differences:
English Japanese
Syllable Mora
Difficult to separate Easy to separate
CCC(V)CCC CCV1(V2)C
Stress-Timed Mora-timing
Tone language
Stress is independent from Stress is independent from
Syllable and dependent syllable
on lexical status Stress comes from variation
in pitch
19 Vowel sounds 5 Vowel sounds
Consonant Clusters permitted Consonant Clusters not Permitted
10 Learner Difficulties
Although many of the world’s languages were designed to reflect pronunciation English is not a direct reflection (Yallop, 1990). The lack of predictabilty regarding pronounciation led Bernard Shaw to complain of a language where ‘fish’ could be spelled ghoti: gh=f (tough), o= /i/ (women), and ti= /ʃ/ (nation) (Pinker, 1994), therefore many English syllables escape any logical method of comprehension causing problems with reading for natives and second language learners. One problem faced only by second language learners is the problems of stress. Incorrect placement of stress is a major problem for foreign learners (Celce-Murcia et al (1996) and is a cause of intelligibility (Roach 2004). As we have seen, selecting the correct syllable to stress in an English word is based on a complicated set of factors which sets English apart from other languages like French or Polish where stress can be decided simply in relation to the syllables of the word i.e. the last syllable or the penultimate syllable. We also looked at Kreidler’s (1989) definition of an important characteristic of stress syllables, in which for example the stressed vowel of `aroma` is the same as the vowel of `row`. This may be of help for learners looking for a stable rule to apply. Celce-Murcia et al (ibid.) ask for syllabuses to determine linguistic priorities to correct any linguistic traits identified in a students first language, so that if a student’s first language is a syllable timed language jazz chants or limericks ‘may have equal priority over excercises that focus on “difficult” segmental contrasts’ (p.324).
Finally, the sequences of English consonant clusters present a challenge for speakers of Japanese that have a stricter pattern of CV order. Celce-Murcia et al (ibid.) have shown that speakers of languages with a simpler syllable structure tend to drop final consonants altogether or change challenging consonant clusters i.e. cold to /kow/. Since the /l/ and /r/ is also a problem for Asian learners they are often ommitted altogether. To overcome this, students can be taught the strategy of cluster reduction known as resyllabification whereby the final consonant of the cluster is moved to the next syllable in order to make the cluster easier to pronounce. Correspondingly, students of Japanese are challenged by a comparable pronunciation transference for raら, tsu [つ] and shi [し]. This is often due to an over-reliance on a Romanji translated alphabet rather than the Hiragana/Katakana alphabet. In the final assessment of the differences in syllables between English and Japanese it must be kept in mind that the goal of language learning/pronunciation must be intelligible – rather than nativelike – pronunciation. This presents a more realistic pedogogical goal as well as respecting the individualism expressed in regional, national and foreign accents.
Bibliography
Blyth, R. H. (1963), A History of Haiku, Volume Two. Tuttle Language Library. Tokyo
Celce-Murcia, M; Brinton, M; Goodwin, J. 1996). Teaching Pronunciation: A reference for Teachers of Other Languages. Cambridge University Press.
Chomsky, N & Halle, M. (1986) The Sound Pattern of English. Harper and Row. New York.
Clark, J & Yallop, C. (1990). An Introduction to Phonetics and Phonology. Second Edition. Blackwell Oxford.
Gilbert, R & Yoneoka, J. (2000). From 5-7-5 to 8-8-8 Haiku Metrics and Issues of Emulation. New Paradigms for Japanese and English Haiku Form
Gimson, A.C. (1962) An Introduction into the pronunciation of English. Edward Arnold Publishers Ltd..
Katamba, F. (1993). Morphology. Palgrave. London.
Kindaichi, H. (1989) The Japanese Language. Tuttle Language Library. Tokyo
Knowles, G. (1990) Patterns of Spoken English: An introduction to English Phonetics. Longman. London
Kono, M. (1997). "Perception and Psychology of Rhythm." Accent, Intonation, Rhythm and Pause. ()
Kreidler, C. (1989). The Pronunciation of English. Basil Blackwell Oxford.
O’Connor, J, D.(1973) Phonetics. Penguin Books.
Pike,K (1943), Phonetics: A critical analysis of phonetic theory an a technique for bthe practical description of sounds. The University of Michigan Press
Pinker, S (1994). The Language Instinct. Penguin Books. UK.
Roach, P. (2004). English Phonetics and Phonology: A practical course. Third Edition. Cambridge University Press.
Shibatani, M.(1990). The Languages of Japan. Cambridge University Press
Stern, H. (2003). Fundamental Concepts of Language Teaching. Oxford University Press.
Tsujimura, N. (1996) An Introduction to Japanese Linguistics. Blackwell Publishers
Internet1: http://ling.mq.edu.au/speech/phonetics/phonoloy/syllable/syll_phonotactic.html
Internet2:Harrington&Cox. http://www.ling.mq.edu.au/speech/phonetics/phonology/syllable/syll_phonotactic.html
Internet 3: 44 phonemes with their common 'sound pattern' representations.
Internet 4: (Gasser , M)
BBC News.(2 December 2005) . Phonics: Strategy but no consensus.
http://news.bbc.co.uk/1/hi/education/4493260.stm
BBC News. (29 June, 2000) One in five UK adults ‘Illiterate’.
http://news.bbc.co.uk/1/hi/uk/811832.stm
Appendix 1
Hirigana
Katakana
The katakana, used to write foreign words, cover exactly the same sounds as hiragana, but are in most cases completely different characters.
Saussure, F. Cours de linguistique générale. 1916
Ladefoged, P. (2001). Vowels and Consonants: An introduction to the Sounds of Languages. Mainden, Mass.: Blackwells
BBC News. One in five UK adults ‘Illiterate’
BBC News Phonics strategy in Schools
Jesperson, Otto. 1922. Language, its nature development and origin. London: George Allen & Unwin.
Whether we talk about stress or accent seems a matter of semantic preference, similarly some authors associate stress with loudness (Trager and Smith 1951: cited in Clark and Yallop 1990) and others are concerned with `stress contours` (Chomsky and Halle 1986).
Celce-Murcia et al (1996) contend that stress placement in English is a product of the English language colourful history. According to Celce-Murcia et al 30 per cent of the vocabulary stems from its Old English origins and Germanic stress patterns, with words of Germanic origins like father, sister and frolic receiving the first syllable as stressed and any words not assimilated to the Germanic pattern receiving a less predictable stress.
Anokutaro O the Buddhas
Sammyaku sambodai no of unexcelled
Hotoke tachi Perfect Enlightenment
Waga tatsu soma ni May you protect the wooded mountains
Myoga arasetamae Where I stand!
(From the Shin-Kokinshu anthology)